Emerging Infectious Diseases [Volume 5 No.3 / May - June 1999] International Editors—Update * Bacterial Resistance to Antimicrobial Agents: Selected Problems in France, 1996 to 1998, H. Aubry-Damon and P. Courvalin Perspectives * The Cost Effectiveness of Vaccinating against Lyme Disease, M.I. Meltzer, D.T. Dennis, and K.A. Orloski * Use of Antimicrobial Growth Promoters in Food Animals and Enterococcus faecium Resistance to Therapeutic Antimicrobial Drugs in Europe, H.C. Wegener, F.M. Aarestrup, L.B. Jensen, A.M. Hammerum, and F. Bager * Bacterial Vaccines and Serotype Replacement: Lessons from Haemophilus influenzae and Prospects for Streptococcus pneumoniae, M. Lipsitch * Iron Loading and Disease Surveillance, E.D. Weinberg Synopses * Human Herpesvirus 6: An Emerging Pathogen, G. Campadelli-Fiume, P. Mirandola, and L. Menotti * Emergence of a Unique Group of Necrotizing Mycobacterial Diseases, K.M. Dobos, F.D. Quinn, D.A. Ashford, C.R. Horsburgh, and C.H. King * Respiratory Diseases among U.S. Military Personnel: Countering Emerging Threats, G.C. Gray, J.D. Callahan, A.W. Hawksworth, C.A. Fisher, and J.C. Gaydos * Q Fever in Bulgaria and Slovakia, V. Serbezov, J. Kazár, V. Novkirishki, N. Gatcheva, E. Kovácová, and V. Voynova * Adhesins as Targets for Vaccine Development, T.M. Wizemann, J.E. Adamou, and S. Langermann Research * Tuberculosis in the Caribbean: Using Spacer Oligonucleotide Typing to Understand Strain Origin and Transmission, C. Sola, A. Devallois, L. Horgen, J. Maïsetti, I. Filliol, E. Legrand, and N. Rastogi * Human Rabies Postexposure Prophylaxis during a Raccoon Rabies Epizootic in New York, 1993 and 1994, J.D. Wyatt, W.H. Barker, N.M. Bennett, and C.A. Hanlon Dispatches * Factory Outbreak of Escherichia coli O157:H7 Infection in Japan, Y. Watanabe, K. Ozasa, J.H. Mermin, P.M. Griffin, K. Masuda, S. Imashuku, and T. Sawada * First Case of Yellow Fever in French Guiana since 1902, J.M. Heraud, D. Hommel, A. Hulin, V. Deubel, J.D. Poveda, J.L. Sarthou, and A. Talarmin * Risk for Rabies Transmission from Encounters with Bats, Colorado, 1977-1996, W.J. Pape, T.D. Fitzsimmons, and R.E. Hoffman * Australian Bat Lyssavirus Infection in a Captive Juvenile Black Flying-Fox, H. Field, B. McCall, and J. Barrett * Bordetella holmesii-Like Organisms Isolated from Massachusetts Patients with Pertussis-Like Symptoms, W.K. Yih, E.A. Silva, J. Ida, N. Harrington, S.M. Lett, and H. George * New Cryptosporidium Genotypes in HIV-Infected Persons, N.J. Pieniazek, F.J. Bornay-Llinares, S.B. Slemenda, A.J. da Silva, I.N.S. Moura, M.J. Arrowood, O. Ditrich, and D.G. Addiss * Fatal Case due to Methicillin-Resistant Staphylococcus aureus Small Colony Variants in an AIDS Patient, H. Seifert, C. von Eiff, and G. Fätkenheuer * Application of Data Mining to Intensive Care Unit Microbiologic Data, S.A. Moser, W.T. Jones, and S.E. Brossette * Sentinel Surveillance for Enterovirus 71, Taiwan, 1998, T.-N. Wu, S.-F. Tsai, S.-F. Li, T.-F. Lee, T.-M. Huang, M.-L. Wang, K.-H. Hsu, and C.-Y. Shen * Chlorine Inactivation of Escherichia coli O157:H7, E.W. Rice, R.M. Clark, and C.H. Johnson * Fulminant Meningococcal Supraglottitis: An Emerging Infectious Syndrome?, E. Schwam and J. Cox * Genetic Evidence of Dobrava Virus in Apodemus agrarius in Hungary, J.J. Scharninghausen, H. Meyer, M. Pfeffer, D.S. Davis, and R.L. Honeycutt * Bacterial Resistance to Ciprofloxacin in Greece: Results from the National Electronic Surveillance System, A.C. Vatopoulos, V. Kalapothaki, Greek Network for the Surveillance of Antimicrobial Resistance, and N.J. Legakis * Emergence of Related Nontoxigenic Corynebacterium diphtheriae Biotype mitis Strains in Western Europe, G. Funke, M. Altwegg, L. Frommelt, and A. von Graevenitz Letters * First Case of Human Ehrlichiosis in Mexico, R.A. Gongóra-Biachi, J. Zavala-Veláquez, C.J. Castro-Sansores, and P.Gonzáles-Martínez * HIV-1 Subtype F in Single and Dual Infections in Puerto Rico: A Potential Sentinel Site for Monitoring Novel Genetic HIV Variants in North America, I. Flores, D. Pieniazek, N. Morán, A. Soler, N. Rodríguez, M. Alegría, M. Vera, L.M. Janini, C.I. Bandea, A. Ramos, M. Rayfield, and Y. Yamamura * Paratyphoid Fever in India: An Emerging Problem, S. Sood, A. Kapil, N. Dash, B.K. Das, V. Goel, and P. Seth * Hepatitis C Virus RNA Viremia in Central Africa, N. Cancré, G. Grésenguet, F.-X. Mbopi-Kéou, A. Kozemaka, A.S. Mohamed, M. Matta, J.-J. Fournel, and L. Bélec * Immunization of Peacekeeping Forces, R. Steffen * Sexually Transmitted Diseases in Ukraine, D.I. Ivanov * Yellow Fever Vaccine, S.C. Arya * Yellow Fever Vaccine—Reply to S. Arya, T.P. Monath, J.A. Giesberg, and E.G. Fierros News and Notes * 54th International Northwestern Conference on Diseases in Nature Communicable to Man _________________________________________________________________________ International Editors—Update _________________________________________________________________________ Update International Editors --------------------------------------------------------------------------- Bacterial Resistance to Antimicrobial Agents: Selected Problems in France, 1996 to 1998 Helene Aubry-Damon* and Patrice Courvalin† *Institut de Veille Sanitaire, Saint-Maurice, France; and †Centre National de Référence des Antibiotiques, Institut Pasteur, Paris, France [picture] Dr. Aubry-Damon is a Surveillance of microbiologist antibiotic resistance specializing in antimicrobial resistance at the in human pathogens has National Institute for Public Health long been performed in Surveillance, Saint-Maurice, France. The France. Existing institute has recently carried out an exploratory surveillance relies on study of a national program for surveillance of national reference antimicrobial resistance in France. Dr. centers dedicated to Aubry-Damon also collaborates closely with the various bacterial French National Reference Center for Antibiotics. genera and on networks (No photo available) of volunteer medical microbiologists, mainly Dr. Courvalin is professor at the Institut in general hospitals Pasteur, where he directs the French National but also in private Reference Center for Antibiotics and heads the laboratories. Regional Antibacterial Agents Unit. Dr. Courvalin data (often initiated specializes in the genetics and biochemistry of at the request of and antibiotic resistance. His research has led to funded by the revision of the description of the natural pharmaceutical dissemination of antibiotic resistance genes. His industry) have been laboratory demonstrated that pathogenic bacteria available since the can promiscuously exchange genetic material early 1950s. Because of conferring antibiotic resistance, documented that the major health conjugation could account for dissemination of problems caused by resistance determinants between phylogenetically antibiotic resistance remote bacterial genera, elucidated the in the last few years, transposition mechanism of conjugative attempts have been made transposons from gram-positive cocci, and to organize a national recently, obtained direct gene transfer from surveillance program bacteria to mammalian cells. similar to those being established in other European countries. Although this update reviews recent data from France, the representativeness of the data has not been assessed. In addition, these are raw data, and their clinical importance remains to be seen; for example, the contribution of bacterial isolates to infection or colonization is, in most instances, unknown. Evolution of Antibiotic Resistance in Hospitals Dissemination of ß-Lactamases in Gram-Negative Bacilli In France, ß-lactamases and fluoroquinolones are the most frequently prescribed antibiotics in Enterobacteriaceae infections. A multicenter study (14 hospitals) across the country analyzed the antibiotic susceptibility of 2,507 and 2,312 consecutive, nonrepetitive enterobacteria responsible for infection in 1996 and 1997, respectively (1,2). Strains were isolated from inpatients (86%) in intensive care (ICU) (12%), surgical (17%), medical (37%), and geriatric (9%) units. The majority of isolates were from urine (71%), pus (9%), and bronchopulmonary specimens (8%). Escherichia coli (64%) was isolated most frequently, mainly in outpatients, whereas Klebsiella spp. and Enterobacteriaceae with inducible ß-lactamases predominated in ICUs. Resistance of E. coli to amoxicillin and cefotaxime was 47% and 0.5%, respectively. In 1997, the frequency of isolates producing an extended-spectrum ß-lactamase varied by species: in Enterobacter aerogenes, 56%; in Klebsiella pneumoniae, 15%; and in E. coli, 0.5%. The incidence differed within and between hospitals. Such strains arise in response to the selective pressure exerted by use of extended-spectrum cephalosporins (3); infections with such strains have also been associated with hospitalization in ICUs. Production of ß-lactamases resistant to ß-lactam-enzyme inhibitor combinations in E. coli was approximately 3.5% (2). Susceptibility to fluoroquinolones was high (66%-97% ciprofloxacin-susceptible) except in E. aerogenes and Serratia marcescens (35%-52% ciprofloxacin-susceptible) (1). The organisms most frequently isolated in ICUs in 1995 belonged to the family Enterobacteriaceae (59%) and Pseudomonas aeruginosa (25%) (4). In 1997, 1,362 nonrepetitive P. aeruginosa (5% of all clinical isolates) were collected in 13 teaching hospitals (5). The lowest rates of susceptibility to ceftazidime (<75%), amikacin (70%), and ciprofloxacin (65%) were observed with serotype 12 (the fourth main serotype). Among penicillinase-producing strains, the percentages of resistance to amikacin and ciprofloxacin were 80% and 93%, respectively; these figures were substantially higher in ß-lactam-resistant P. aeruginosa than in susceptible strains. Spread of Methicillin-Resistant and Gentamicin-Susceptible Staphylococcus aureus Methicillin-resistant S. aureus (MRSA) is one of the most frequent nosocomial pathogens in France as in the rest of the world. Surveys conducted in hospitals in Paris and surroundings found that MRSA decreased from 42% in 1992 to 37% in 1997 and that the incidence of MRSA colonization-infection (approximately 0.65 per 100 admissions) also decreased after national recommendations against dissemination of multidrug-resistant bacteria were implemented (6,7). However, a survey of 26 geographically representative hospitals found that the incidence of gentamicin-susceptible MRSA progressively increased because of the presence of a predominant clone (H. Lelièvre, G. Lina, M.E. Jones, et al., unpub. obs.). The epidemiologic situation in France is complex. The endemic aminoglycoside-resistant MRSA strain persisted while new clones became endemic in hospitals, perhaps after changes in the use of aminoglycosides (decrease of gentamicin and increase of amikacin consumption) (8). The first vancomycin-intermediate S. aureus was isolated in a French hospital in 1995; no other cases of MRSA with reduced susceptibility to vancomycin have been reported (9). Most Frequent Macrolide-Resistance Mechanisms among Staphylococci Over 3 weeks in 1995, 607 staphylococci were collected in 32 hospitals (10). Of these, 45.5% of the S. aureus and 54% of the coagulase-negative staphylococci were resistant to methicillin, and 71.5% of MRSA were resistant to macrolides. Of these MRSA strains, 75% were constitutively resistant, whereas 76% of MSSA were inducibly resistant. A similar distribution (61% vs. 27.5%) was observed among coagulase-negative staphylococci. Resistance to at least one of the macrolide, lincosamide, and streptogramin antibiotics (88%) was due to the presence of the ermA and ermC genes, which confer resistance by modifying the ribosomal target. The ermA gene was more common in MRSA (57.6%) than in MSSA (5.6%), where ermC was predominant (20.1%). ermC was also common among methicillin-susceptible coagulase-negative staphylococci (14%). Only a few strains had the ermB gene, which is found in animal strains. Macrolide resistance by efflux due to acquisition of the msrA gene was more prevalent in coagulase-negative staphylococci (14.6%) than in S. aureus (2.1%). The incidence of lincomycin-resistant but macrolide- and streptogramin-susceptible staphylococci was low: 0.2% in S. aureus and 4.6% in Staphylococcus epidermidis (11). The prevalence of pristinamycin-resistant (and also most probably quinupristin/dalfopristin-resistant) strains remained low because of the low incidence of resistance to streptogramins type A (pristinamycin has limited use in France). Dissemination of Resistance in the Community Effect of Antibiotics on Oropharyngeal Flora Antibiotic Resistance in Streptococcus pneumoniae The National Reference Center for Pneumococci determined the susceptibility to antibiotics of 2,837 S. pneumoniae isolated in 1997. The incidence of S. pneumoniae with reduced susceptibility to penicillin G increased from 3.8% in 1987 to 48% in 1997 (12). Whereas 53% of all strains were resistant to macrolides, 80% of penicillin-resistant strains were macrolide-resistant; 15% of all strains (versus 51% of penicillin-resistant strains) were resistant to tetracycline, and 10% (versus 66%), respectively, were resistant to trimethoprim-sulfamethoxazole (F. Goldstein et al., unpub. data). According to a 1997 survey of 18 regional laboratories in France (11,757 strains collected), 27% had intermediate levels of resistance to penicillin G, and 13.5% were fully resistant. The rates varied considerably by region (highest in southwest and central France), age (highest [37.4% penicillin G-intermediate and 21.5% resistant] in children <16 years old), and specimen source (highest in middle ear and sinus specimens) (13). Resistance to Macrolides in ß-Hemolytic Streptococci In 1995, a national survey in 98 hospitals of invasive infections due to Streptococcus pyogenes found that 5.2% to 9.8% of the strains isolated from blood were erythromycin-resistant (A. Bouvet, pers. comm.) (14). Vaccination against and Resistance in Haemophilus influenzae Type b To monitor the trends in H. influenzae meningitidis and the prevalence of resistance, the National Reference Center conducted a survey of approximately 80 hospitals (15). Since vaccination for Hib invasive infections began in 1993, the percentage of capsulated isolates has decreased 5% per year. Moreover, resistance to antimicrobial drugs decreased among Hib and increased among noncapsulated strains isolated from upper and lower respiratory tract infections. The percentage of ß-lactamase-producing H. influenzae increased progressively from 22% in 1992 to 35% in 1997, with a similar evolution for kanamycin resistance. Tetracycline and chloramphenicol resistance remained stable in 1997—less than 10% and 2%, respectively (15,16). Antibiotic Resistance in Neisseria meningitidis Meningococcal resistance to antibiotics is emerging in France. The incidence of N. meningitidis with reduced susceptibility to penicillin G (MICs from 0.125 mg/L to 1 mg/L) increased from less than 1% in 1991 to 18% in 1996 (17,18). The strains belonged to various serogroups; most belonged to serogroup B, none produced a ß-lactamase, and all were susceptible to cefotaxime and ceftriaxone. Resistance to rifampin, used for prophylaxis of secondary cases in France, remained low (0.02% in 1996). Effect of Antibiotics on Digestive Flora Antimicrobial Resistance in Helicobacter pylori Susceptibility testing of H. pylori from 535 patients with a positive CLO test was performed in 1997 (19). Depending on the method, the percentages of clarithromycin resistance (disk-agar diffusion or MIC determination by agar dilution) and metronidazole resistance (breakpoint method at 8 mg/L or MIC determination) varied from 14.3% (95% confidence interval [CI] 11.5-17.6) to 14.0% (95% CI 11.2-17.3) and from 30.5% (95% CI 25.6-34.5) to 23.6% (95% CI 20.1-27.5), for the two antibiotics, respectively. No resistance to amoxicillin was observed. Fluoroquinolone Resistance in Campylobacter and Salmonella Hadar The evolution of antimicrobial resistance in Campylobacter jejuni and C. coli is worrisome. Between 1986 and 1997, 2,713 strains of C. jejuni (68% of total Campylobacter isolates) were isolated from stool (94%) and blood (4%) and studied (20). Between 1993 and 1997, fluoroquinolone resistance increased from 7.4% to 32% in C. jejuni and from 11.8% to 52% in C. coli. The high resistance rate to quinolones makes them ineffective in therapy of Campylobacter infections. These resistance rates are similar to those in other countries (e.g., Spain, the United Kingdom) (21,22). However, the prevalence of macrolide-resistant strains remains low (3.6%). The high incidence of multidrug-resistant Salmonella Typhimurium DT104 (12 atypical), with 82% resistance to ampicillin, streptomycin, sulphonamide, tetracycline, and chloramphenicol, is the most serious epidemiologic problem of the last decade in France (23). The incidence of Salmonella Hadar is increasing, and the percentages of amoxicillin- and fluoroquinolone-resistant strains in 1997 were 72% and 75%, respectively. Fluoroquinolone resistance had not been observed before 1987 in France, Spain, and the United Kingdom. This was before concomitant introduction of ciprofloxacin into clinical use and enrofloxacin into veterinary use (in particular in the poultry industry) in the late 1980s. More than 50% of C. jejuni and S. Hadar, the most frequent serotype associated with poultry, are now fluoroquinolone-resistant in these countries. The situation is different in Sweden, where fluoroquinolones are not readily available. Therefore, guidelines for the prudent use of antibiotics (in prophylaxis or therapy) should be developed that respect the indigenous flora of humans and animals. New Types of Resistance in Enterococci The increase in the incidence of glycopeptide-resistant enterococci (GRE) isolated from hospitalized patients throughout the United States has not been observed in France. A multicenter study in 1993 showed a very low incidence of GRE: 0.2% among 251 enterococcal clinical isolates and 7.5% among Enterococcus faecium (24). Study of 24 ICUs in 1994 determined that the prevalence of GRE colonization in patients' fecal flora was approximately 2%, 30% of which had been present at admission. No nosocomial infection due to GRE was observed (25). GRE have been identified in human food of animal origin (40% of GRE were isolated from uncooked meat) in a French study conducted in military cafeterias in 1997 (26). Thus, food may represent a major source of human colonization with GRE in France. GRE strains isolated in France were also resistant to ampicillin, tetracycline, and macrolides. However, the percentage of high resistance levels to gentamicin among GRE was comparable to that among glycopeptide-susceptible enterococci. Antibiotic Resistance in Bacteroides fragilis Studies of antibiotic resistance in anaerobic pathogens indicate stability of resistance to carbapenems (imipenem) and nitroimidazole antibiotics (27,28). In 1998, fewer than 2% of all B. fragilis from 39 hospitals were resistant to metronidazole (MICs >8 mg/L), and the number of imipenem-resistant strains remained low. However, this gene reservoir requires surveillance of resistance in B. fragilis infections because of the use of these antibiotics in therapy. Other Bacteria Antibiotic Resistance in Neisseria gonorrhoeae The number of N. gonorrhoeae strains identified by the National Reference Centre for Sexually Transmitted Diseases fell sharply from 1986 to 1990 (by 81%) and more slowly from 1990 until 1999 (by 55%) (29). The number of anorectal gonococcal infections reached a plateau from 1995 to 1997 but increased again in 1998, mostly in the Paris/Ile-de-France region (V. Goulet, P. Sednaoui, et al., unpub. data). An increasing percentage of N. gonorrhoeae displayed diminished sensitivity to penicillin G and to tetracycline. In 1997, 15% and 30% of N. gonorrhoeae were resistant to penicillin G (MIC >/= 2 mg/L) and tetracycline (MIC 2 >/= mg/L and < 16 mg/L), respectively, by chromosomal mutation. In contrast, the percentage of strains with plasmid-mediated resistance to penicillins and tetracycline has remained stable at approximately 15% since 1994. No ceftriaxone, spectinomycin, or ciprofloxacin resistance was found until 1997, when the first ciprofloxacin-resistant strains (MIC=1 mg/L) were isolated. Conclusions Antibiotic resistance trends in France are for the most part similar to trends in other European countries but with some peculiarities. For instance, fluoroquinolone resistance in Salmonella spp. and Campylobacter spp. is a problem throughout Europe. However, methicillin resistance in Staphylococcus is more common in France than in the Scandinavian countries, although it has started to decrease because of reinforcement of hygiene measures since 1992. Also, heavy use of 16-membered macrolides has selected for resistance in gram-positive cocci by ribosomal modification rather than by efflux. In pneumococci, decreased susceptibility to penicillins is as common in France as in Spain, but the incidence of resistance to macrolides is the highest in Europe. The public health problems caused in France by bacterial resistance to antibiotics are clearly distinct from those in North America. The incidence of enterobacteria producing extended-spectrum ß-lactamases and glycopeptide-resistant enterococci remains rather low in France, as in most other European countries. In the United States, the high incidence of nosocomial GRE infections is probably caused by the heavy nosocomial use of vancomycin, particularly in hematology wards and for the prevention of colitis due to Clostridium difficile. In contrast, no intestinal carriage of such strains is found in the general population. The situation in Europe mirrors that in the United States. In Europe, the prevalence of nosocomial GRE infections remains low, but colonization of the population is substantial, possibly because of the use of a vancomycinlike antibiotic (avoparcin) as an animal food additive. This example stresses the need for a multidisciplinary approach to surveillance of bacterial resistance to antibiotics. --------------------------------------------------------------------------- Address for correspondence: P. Courvalin, Unité des Agents Antibactériens, Institut Pasteur, 28, rue du Dr. Roux, 75724 Paris Cedex 15, France; fax: 33-1-45-68-83-19; e-mail: pcourval@pasteur.fr. References 1. Chardon H, Nicolas-Chanoine MH, Sirot J, and le Groupe d'Etude Multicentrique. Evaluation de la sensibilité des Enterobacteriaceae aux b-lactamines et aux fluoroquinolones: Résultats d'une enquête multicentrique en 1996 et 1997. Proceedings of the 18th Interdisciplinary Meeting on Anti-Infectious Chemotherapy; 1998 Dec 3-4; Paris, France. p. 129. 2. Nicolas-Chanoine MH, Sirot J, and le Groupe d'Etude Multicentrique. Caractérisation et distribution des mécanismes de résistance aux b-lactamines parmi les entérobactéries: résultats d'une enquête multicentrique en 1996. Proceedings of the 17th Interdisciplinary Meeting on Anti-Infectious Chemotherapy; 1997 Dec 4-5; Paris, France. p. 251. 3. Brun-Buisson C, Legrand P, Philippon A, Montravers F, Ansquer H, Duval J, et al. Transferable enzymatic resistance to third-generation cephalosporins during nosocomial outbreak of multiresistant Klebsiella pneumoniae. Lancet 1987;2:302-6. 4. Jarlier V, Fosse T, Philippon A, for the ICU Study Group. Antibiotic susceptibility in aerobic gram-positive bacilli isolated in intensive care units in 39 French teaching hospitals (ICU study). Intensive Care Med 1996;22:1057-65. 5. Cavallo JD, Leblanc F, Thabaut A, Groupe d'Etude de la Résistance de P. aeruginosa aux ß lactamines. Susceptibility of Pseudomonas aeruginosa to nine antimicrobials: a 1997 French multicenter hospital survey. Proceedings of the 38th Interscience Conference on Antimicrobial Agents and Chemotherapy; 1998 Sep 24-27; San Diego, California. Washington: American Society for Microbiology; 1998. p. 191. 6. The College de Bacteriologie-Virologie-Hygiene du Centre Hospitalier Universitaire de Paris. Surveillance des staphylocoques dorés et klebsielles multirésistants à l'Assistance Publique-Hôpitaux de Paris, 1993-1996. Bulletin Epidémiologique Hebdomadaire 1998;10:41-3. 7. Reseau de Microbiologie du C.CLIN Paris Nord et le Groupe de Microbiologistes d'Ile-de-France. Surveillance des bactéries multirésistantes a partir du laboratoire. Bulletin du Centre de Coordination de la Lutte contre les Infections Nosocomiales, Paris-Nord 1998;11:4-5. 8. Aubry-Damon H, Legrand P, Brun-Buisson C, Astier A, Soussy CJ, Leclercq R. Reemergence of gentamicin-susceptible strains of methicillin-resistant Staphylococcus aureus: role of an infection control program and changes in aminoglycoside use. Clin Infect Dis 1998;25:647-53. 9. Ploy MC, Grelaud C, Martin C, de Lumley L, Denis F. First clinical isolate of vancomycin-intermediate Staphylococcus aureus in French hospital. Lancet 1998;351:1212. 10. Lina G, Quaglia A, Reverdy ME, Leclercq R, Vandenesch F, Etienne J. Distribution of genes encoding resistance to macrolides, lincosamides and streptogramins among staphylococci. Antimicrob Agents Chemother. In press 1999. 11. Leclercq R, Brisson-Noel A, Duval J, Courvalin P. Phenotypic expression and gene heterogeneity of lincosamide inactivation in Staphyloccus spp. Antimicrob Agents Chemother 1991;31:1887-91. 12. Geslin P. National Reference Center for Pneumococci. France, Final Activity Report 1997. 13. Roussel-Delvallez M, Weber M, Maugein J, Thierry J, Laurans G, Fosse T, et al. Résistance du pneumocoque aux antibiotiques en 1997: résultats de 18 observatoires régionaux. Bulletin Epidémiologique Annuel 1998 report. France: National Institute for Public Health Surveillance. In press 1999. 14. Varon E, Havlickova H, Pitman C, Sarr A, Muller-Alouf H, Coignard S, et al. Comparison of invasive (septicemic) and non invasive strains of group A streptococci isolated during a one-year national survey in France. Adv Exp Med Biol 1997;418:83-5. 15. Dabernat H. Donnees de surveillance du Centre National de Référence des Haemophilus influenzae: avant et apres la vaccination fr. Bulletin Epidémiologique Annuel 1998 report. France: National Institute for Public Health Surveillance. In press 1999. 16. Dabernat H, Delmas C. Activité du Centre National de Référence des Haemophilus influenzae, années 1996-1997: le déclin du type b. Medecine et Maladies Infectieuses 1998;28:940-6. 17. Guibourdenche M, Lambert T, Courvalin P, Riou JY. Epidemiological survey of Neisseria meningitidis susceptibility to penicillin G in France. Pathol Biol 1997;45:729-36. 18. Struillou L, Chamoux C, Berranger C, Chouillet AM, Riou JY, Raffi F. Rapid emergence of meningococci with reduced susceptibility to penicillin in France: the need for vigilance in meningitidis treatment. Clin Microbiol Infect 1998;4:661-2. 19. Broutet N, Guillon F, Sauty E, Lethuaire D, Megraud F. Survey of the in vitro susceptibility of Helicobacter pylori to antibiotics in France. Gut 1998;43:All. 20. Megraud F. Les infections a Campylobacter en France 1986-1997, le Centre National de Référence des infections à Campylobacter. Bulletin Epidémiologique Annuel 1998 report. France: National Institute for Public Health Surveillance. In press 1999. 21. Gaunt PN, Piddock LJV. Ciprofloxacin-resistant Campylobacter spp. in human—san epidemiologic and laboratory study. J Antimicrob Chemother 1996;37:747-57. 22. Reina J, Alomar P. Fluoroquinolone resistance in thermophilic Campylobacter spp. Lancet 1990;336:186. 23. Breuil J, Armand-Lefevre L, Casin I, Dublanchet A, Collatz E and The College de Bacteriologie-Virologie-Hygiene des Hôpitaux Généraux Français. Surveillance de la sensibilité aux antibiotiques des salmonelles et shigelles isolées dans 77 hôpitaux français. Bulletin Epidémiologique Hebdomadaire 1998;51:219-21. 24. Schmit JL, Leclercq R, Scheimberg A, Landauer D. Approche épidemiologique et clinique des entérocoques: résultat d'une enquête. Medecine et Maladies Infectieuses 1994;24S:141-8. 25. Boisivon A, Thibault M, Leclercq R, and The College de Bacteriologie-Virologie-Hygiene des Hôpitaux Généraux Français. Colonization by vancomycin-resistant enterococci of the intestinal tract of patients in intensive care units from French general hospitals. Clin Microb Infect 1997;3:175-9. 26. Perrier-Gros-Claude JD, Courrier PL, Breard JM, Vignot JL, Masseront T, Garin D, et al. Entérocoques résistants aux glycopeptides dans les viandes. Bulletin Epidémiologique Hebdomadaire1998;12:50-1. 27. Breuil J, Podglajen I, Collatz E. Susceptibility testing of anaerobic pathogens: rational and results. Proceedings of the 38th Interscience Conference on Antimicrobial Agents and Chemotherapy; 1998 Sep 24-27; San Diego, California. Washington: American Society for Microbiology; 1998. p. 636. 28. Reysset G, Trinh S, Carlier JP, Sebald M. Bases génétiques de la résistance aux 5-nitroimidazoles des Bacteroides spp. Medecine et Maladies Infectieuses 1996;26 Suppl:1-7. 29. National network on gonococcal infections. Les gonococcies en France en 1997, le réseau RENAGO. Bulletin Epidémiologique Annuel 1998 report. France: National Institute for Public Health Surveillance. In press 1999. _________________________________________________________________________ Perspectives _________________________________________________________________________ Perspectives The Cost Effectiveness of Vaccinating against Lyme Disease Martin I. Meltzer, David T. Dennis, and Kathleen A. Orloski Centers for Disease Control and Prevention, Atlanta, Georgia, USA --------------------------------------------------------------------------- To determine the cost effectiveness of vaccinating against Lyme disease, we used a decision tree to examine the impact on society of six key components. The main measure of outcome was the cost per case averted. Assuming a 0.80 probability of diagnosing and treating early Lyme disease, a 0.005 probability of contracting Lyme disease, and a vaccination cost of $50 per year, the mean cost of vaccination per case averted was $4,466. When we increased the probability of contracting Lyme disease to 0.03 and the cost of vaccination to $100 per year, the mean net savings per case averted was $3,377. Since few communities have average annual incidences of Lyme disease >0.005, economic benefits will be greatest when vaccination is used on the basis of individual risk, specifically, in persons whose probability of contracting Lyme disease is >/=0.01. Lyme disease, caused by infection with Borrelia burgdorferi, is the most common tick-borne disease in the United States and Europe (1-3). In the United States, the disease has spread slowly, and the number of cases in disease-endemic areas has increased (4-6). Most Lyme disease patients become infected with B. burgdorferi near their homes, while engaged in property maintenance, recreation, and relaxation (7). Occupational and recreational activities away from home may also pose a risk (8). Lyme disease prevention based primarily on avoidance of tick bites, use of repellants, early detection and removal of attached ticks, and tick control has not substantially reduced disease incidence (4-6). Therefore, preventive vaccines have been of considerable interest. Results of randomized and blinded phase-III field trials with recombinant B. burgdorferi outer surface protein A (rOspA) vaccines indicate that they are safe and efficacious (9,10). On December 21, 1998, the U.S. Food and Drug Administration licensed one of the vaccines (LYMErix, SmithKline Beecham Biologicals, Reixensart, Belgium) for use in the United States (11). We present the results of an analytic model that evaluates the cost effectiveness of using a vaccine to protect against Lyme disease in the United States. The Model Using a computer-based spreadsheet (Excel 5.0 for Windows, Microsoft), we constructed a decision tree (12) to evaluate the cost per case averted (cost effectiveness) to society of vaccinating against Lyme disease (Figure 1). Many data needed to determine the cost effectiveness of vaccinating against Lyme disease are unvalidated, unavailable, or available only from very small databases. Thus, rather than calculate a single estimate of cost per case averted, we examined the effect of combinations of six inputs: cost of vaccination; annual probability of contracting Lyme disease; costs of successfully treating either early symptoms of Lyme disease or one of three sequelae (cardiovascular, neurologic, arthritic); probability of diagnosing and treating early symptoms; probability of sequelae due to early infection; probability of sequelae due to late, disseminated infection. Mathematically, we examined the effect [Fig] of altering the values of the inputs by using specialized computer software Fig 1. Decision tree to model (@Risk, Palisade Corp., Newfield, NY) the cost effectiveness of (13) that employs Monte Carlo methods vaccinating a person against (14-16). To use these methods, the Lyme disease. researcher defines probability distributions for selected inputs by using available data and enters them into the computer program. For each iteration of the program, an algorithm selects input values from the probability distributions, calculates the net result (here, the cost per case averted), and stores that result. After all iterations (typically 1,000 to 5,000) are completed, the program produces a probability-based distribution of the net result, which can then be used to report statistics such as mean, median, and 5th and 95th percentiles. Cost Effectiveness Formula The formula used to calculate the cost per case of Lyme disease averted was as follows: Cost per case averted = ($ of vacc+$of LD with vacc-$ of LD without vacc)/ (Prob LD without vacc-Prob LD with vacc) where $ = cost; vacc. = vaccination; LD = Lyme disease; and prob. = probability. The numerator is the cost of vaccination less any savings resulting from the reduced probability of contracting the disease (decreased incidence) due to vaccination. If the vaccine is not 100% effective in preventing Lyme disease (i.e., if the term Prob. LD with vacc. >0), treatment costs may still be incurred after vaccination. The cost of a case of Lyme disease is the weighted average cost of all health outcomes (Figure 1), where the weights are the probabilities of those outcomes 12). The denominator reflects the change in the probability of Lyme disease due to vaccination. Vaccine Timeline Although experiments have shown that a Lyme disease vaccine using rOspA is safe and immunogenic in both animals and humans (17-23), no data have been published concerning the decrease in antibody levels over more than 20 months (9). Phase-III vaccine field trials used a 0-, 1-, and 12-month immunization schedule, and antibody levels dropped almost 10-fold between the month after the second dose and just before the third dose at month 12 (9). The third dose at month 12 boosted antibodies to levels higher than measured at month 2, but these declined by half by month 20 (9). We assumed, therefore, that an annual booster dose would be required and that the cost-effectiveness model would be repeated annually. When calculating annual benefits, however, we included the discounted savings of preventing Lyme disease that may generate multiyear sequelae. Lyme Disease Symptoms and Sequelae The most common symptoms of infection with B. burgdorferi can be categorized as early localized disease (stage I); early disseminated disease (stage II); and later stage sequelae of disseminated infection-(stage III) (24). Stages I and II correspond to the branches labeled "Recognize early LD? Yes" in Figure 1, and stage III corresponds to the branches labeled "Recognize early LD? No." Most early symptoms of Lyme disease respond promptly and completely to short courses of oral antibiotics (25-27). Later-stage sequelae, however, may require costly, more prolonged treatment, sometimes repeated courses of treatment using intravenous cephalosporins, and may not be completely eliminated (28). If a person, vaccinated or unvaccinated, contracts Lyme disease, the model allows for one of four possible categories of outcomes (Figure 1) (29-31): cardiovascular sequelae (e.g., high-grade atrioventricular blocks); neurologic sequelae (e.g., isolated cranial nerve palsy, meningitis); arthritic or rheumatologic/musculoskeletal sequelae (e.g., episodic oligoarticular arthritis, arthralgia); and case resolved (after a course of an oral antibiotic such as doxycycline) with no further complications. The disseminated stages of Lyme disease may be manifested weeks to months after infection (24). However, few data concerning the duration of such sequelae are available. One study, for example, involving 38 patients showed that their long-term clinical sequelae lasted a mean of 6.2 years from onset of disease (32). The use of health-care resources, however, by those patients during that time was not reported. We assumed that cardiovascular sequelae would be treated and resolved in an average of 1 year and that late neurologic and arthritic sequelae would both take an average of 11 years to diagnose and satisfactorily treat to full resolution (initial year of diagnosis and treatment plus 10 years of additional treatment). These assignments of average time are arbitrary and longer than any published average, which maximizes estimated economic benefits of using a vaccine. Probabilities We selected three probabilities (0.005, 0.01, and 0.03) of contracting Lyme disease (Table 1) on the basis of data concerning disease incidence in Lyme disease-endemic areas ((33-36); the probability of 0.03 is among the highest reported. (Before the risk for Lyme disease was widely recognized, a one-time annual incidence of 10% was reported in a community of 190 people living next to an open nature preserve [37].) Vaccine efficacy in preventing Lyme disease was 50% (95% confidence intervals [CI]: 14% to 71%) after the first two doses and 78% (95% CI: 59% to 88%) after three doses (9,11). Table 1. Probabilities and their statistical distributions --------------------------------------------------------------------------- Item Values Type of distribution(sup a) --------------------------------------------------------------------------- Probability of contracting LD(sup b) 0.005, 0.01, 0.03 Fixed intervals (sup c) Effectiveness of vaccine 0.85 Fixed Probability of early Fixed intervals (sup c,d) detection of LD 0.6 - 0.9 Probability of sequelae (sup e) if detect LD early Cardiac 0 - 0.01 Uniform (sup f) Neurologic 0 - 0.02 Uniform (sup f) Arthritic 0.02-0.05-0.07 Triangular (sup g) Case resolved Residual (sup h) N/A Probability of sequelae if do not detect LD early Cardiac 0.02-0.03-0.06 Triangular Neurologic 0.02-0.15-0.17 Triangular Arthritic 0.5-0.6-0.62 Triangular Case resolved Residual (sup h) N/A --------------------------------------------------------------------------- (sup a) Statistical distribution used in Monte Carlo simulations (14-16). (sup b) LD = Lyme disease. (sup c) Iterations are run by using different combinations of the probabilities of infection and cost of treatment (Table 2). (sup d) The interval between the minimum and the maximum is divided into 0.1 increments. (sup e) See text for description of sequelae. (sup f) Uniform distribution implies that there is an equal chance that any number between, and including, the minimum and maximum will be used for a given iteration. (sup g) Triangular distribution is defined by points of minimum, most likely, and maximum. (sup h)The probability of an LD case being successfully resolved (i.e., no further sequelae) is 1 - (sum of the probabilities of cardiac + neurologic + arthritic symptoms). We assumed Lyme disease vaccine to be 85% effective, which is near the upper end of the 95% confidence limits and thus maximizes estimated economic benefits. We selected 0.6 to 0.9 as the range of probability of early diagnosis and treatment on the basis of a study on the economic cost of Lyme disease, which included data from an expert panel (38). For the Monte Carlo simulations (14-16), we constructed the distributions describing the probabilities of having one of the three sequelae (due to either early or late disseminated disease) using data from the previously mentioned expert panel (Table 1) (38). The distributions describing cardiac and neurologic complications associated with early Lyme disease are uniform, defined by using minimum and maximum values (39) and reflecting the uncertainty regarding a most likely value (38). All other distributions are triangular (39), with minimum, most likely, and maximum values (Table 1). Vaccination Costs Although a Lyme disease vaccine has been licensed (11), data are not available on the actual cost of vaccination, which includes costs of the vaccine, its administration, time spent in receiving the vaccine, travel, and treatment of adverse side-effects of vaccination. To allow for variation caused by variables such as location of provider, type of provider, and type of third-party payer, we estimated cost effectiveness by using three costs: $50 per person per year, $100 per person per year, and $200 per person per year. Few data are available on the costs of treating a case of Lyme disease; only one study (29) has documented the charges in 1989 dollars associated with some sequelae. To adjust charges reported in that study to 1996 prices, we multiplied the charges by a factor of 1.528 (medical care component of the consumer price index) (40). These 1996 prices, however, reflected health-care charges paid by health insurance companies and not necessarily actual economic costs (41,42). Thus, to reflect economic costs, the adjusted prices were multiplied by cost-to-charge factor (the weighted average of the urban and rural hospital cost-to-charge ratios used by the U.S. Federal Health Care Finance Administration [43]) of 0.53. Data describing indirect costs, particularly lost productivity, associated with sequelae were unavailable. We therefore assumed that Lyme disease–related cardiac sequelae would cause 14 days of lost productivity, and neurologic and arthritic sequelae would each cause 21 days of lost productivity per year. Each day of lost productivity was valued at $100 (the average income of a workday [1990 dollars inflated to 1996 values] weighted by the age and sex composition of the U.S. workforce) (44). Because we assumed that late-stage neurologic and arthritic complications may take up to 11 years to completely resolve, the 1-year cost estimates for treating these sequelae were replicated over 11 years and then discounted at 3% to the base year (Table 2). Table 2: Costs of treating one case of Lyme We also disease and the sequelae due to early and late altered disseminated disease the estimate ----------------------------------------------------------------- of Magid Cost/ Length Total et al. year of treat- costs(sup a)(29) of Item ($) ment ($) charges for ----------------------------------------------------------------- resolving Case resolved: no sequelae a case Antibiotics 14 of Lyme disease Office visits (2) 50 without Laboratory tests 35 complications 5 hrs lost work time 62 by doubling Total 161 2-3 wks 161 the Sequelae (sup b)due to early and number late disseminated disease of office Cardiac-direct(sup c) 5,445 visits Cardiac-indirect (sup d) 1,400 to two ($25 Cardiac-total 6,845 /= 10 in either test were considered serologically positive. Acute Q-fever diagnosis was based on seroconversion or fourfold rise of antibody titers. Similar results were observed when testing human sera from the serum bank. Of 224 randomly chosen sera collected in nine localities (Gabrovo from the North, Razgrad and Dobrich from the Northeast, Sofia from the West, Blagoyevgrad from the Southwest, and Stara Zagora, Pazardzhik, Haskovo, and Sliven from the South of Bulgaria), 87 (38%) reacted positively with phase II C. burnetii in MA or MIF tests. Serologic positivity varied from 6% in Sliven to 60% in Blagoyevgrad, except for Razgrad, where none were positive (however, only four sera were tested). Chronic Q-fever cases manifesting as endocarditis were confirmed serologically by high titers from 640 to 1 ml of phase I- and phase II- C. burnetii immunoglobulin (Ig)G antibodies in MIF, by demonstration of specific immunofluorescence in the cuts of aortal valves, and by C. burnetii isolation from the replaced prosthesis in three patients before 1990 (23). Two additional cases of Q-fever endocarditis were diagnosed serologically by MA and MIF tests from 1996 to 1997. Antibodies to phase II C. burnetii by MA were found in 16 of 18 aborting women with titers of 10 to 320, which indicates the possibility of acute Q-fever infection during pregnancy. In two cases, in paired sera collected in 23-day intervals, a shift from the titer of 160 to seronegativity (titer <10) was observed. Even though abortion tissues were not cultured or tested for C. burnetii, these findings deserve further study, since the possible adverse effects of C. burnetii infection during pregnancy has also been suggested by other authors (24). Q Fever in Slovakia In Slovakia, Q fever has been known since 1954 when outbreaks occurred among agricultural workers who contracted the infection from sheep imported from Romania and among workers of a textile plant who were exposed to contaminated imported cotton (6). From that time until the 1980s, the waves of epizootics and small epidemics appeared in factories processing cotton, wool, and hides from Mongolia and China, in a sheep farm with imported breeding rams from England, and in various agricultural premises often connected with excursions of workers to cattle or sheep farms in which C. burnetii infections could have occurred. Veterinary and serologically uncontrolled movement of cattle within the country also contributed to the establishment of domestic coxiellosis (25). Some areas of the southern part of central Slovakia became a natural focus of Q fever, with the D. marginatus tick as the main vector of C. burnetii (6). Since the 1980s, only sporadic cases of Q fever have been reported from different parts of the country, though almost 3% of approximately 7,000 ticks collected in all districts of Slovakia were found (by the hemocyte test) to harbor C. burnetii, and attempts to recover C. burnetii from pooled positive ticks resulted in the isolation of 10 virulent C. burnetii strains from five ticks, mostly I. ricinus species (26). On the other hand, C. burnetii strains isolated from cow milk were of lower virulence for guinea pigs and mice (Kovácová et al., submitted for publication). Circulation of such low virulent strains among livestock and large-scale vaccination of cattle by inactivated phase I-C. burnetii corpuscular vaccine (one subcutaneous dose consisting of 500 µg of highly purified C. burnetii cells) carried out in the 1970s and 1980s, together with improved veterinary control of domestic animal transport within the country, could explain a decrease in the occurrence of human Q fever in Slovakia. This explanation is supported by results of a serologic survey (carried out from 1989 to 1996) for Q-fever antibodies in groups of farmers or in patients with suspected C. burnetii infection. Of 21,197 human sera tested, 655 (3%) reacted with phase II- C. burnetii antigen in the CF test (until 1992) or (later on) in enzyme-linked immunosorbent assay (ELISA). Acute Q fever (as individual or clustered cases) was diagnosed in 23 sera, not including 113 from the Q-fever epidemic discussed below, on the basis of seroconversion or IgM antibody detection. During the same period, phase-II C. burnetii antibodies were detected in 11% of cattle and in 3% each of sheep and goats. Improved veterinary control of domestic animal transport within the country, however, cannot exclude the possibility of introducing C. burnetii infection through imported domestic animals or raw materials not tested properly. The use of CF, which is much less sensitive to Q-fever antibodies than other serologic tests (e.g., ELISA [27]), to screen the goats imported to Slovakia from Bulgaria is unsatisfactory, as confirmed in 1993 by the largest reported Q-fever epidemic in Slovakia (28). The epidemic started suddenly during the spring as an outbreak of respiratory infection in inhabitants of a village in West Slovakia. A total of 113 persons were affected from the beginning of March until May 18, as confirmed serologically (seroconversion, detection of IgM antibodies, and high phase-II antibody titers, respectively) by CF, MA, MIF, and ELISA. Of 42 patients admitted to the hospital, 33 had atypical pneumonia (diagnosed by X-ray examination), and 27 had hepatic involvement (diagnosed on the basis of the increased values of liver transaminases). As many as 103 were male patients who used to visit the local pub, in which they contracted infection by the aerosol created from the heavily contaminated garments of boys tending aborting goats. C. burnetii infection in incriminated goats was confirmed serologically (46 of 216 goat sera tested were positive by ELISA) and by seroconversion in mice inoculated with spleen, lung, and liver suspension from an aborted kid. In contrast to the situation in Bulgaria, a 4-year follow-up of patients from this Q-fever epidemic did not result in clinical or serologic confirmation of any chronic form of the disease (Kovácová et al., submitted for publication). In addition, evidence of chronic Q fever was obtained neither in the serologic survey carried out in Slovakia from 1989 to 1996, nor in testing of more than 200 patients with chronic cardiovascular disease (some of them exposed to C. burnetii infection through their work). Similarly, observation of patients from other Q-fever epidemics (including those with 98 cases in a cotton-processing plant in nearby Southern Moravia in 1980 [29]) was also negative. Whether this can be explained by C. burnetii strains of different virulence circulating in Bulgaria and Slovakia, respectively, remains to be seen, though in the latest Q-fever epidemic in West Slovakia, Bulgarian C. burnetii strains were presumably involved. However, whereas C. burnetii strains of tick and domestic animals origin isolated in Slovakia may differ, no data are known on the virulence of Bulgarian strains. A total number of human C. burnetii cases can also be important. Whereas in Bulgaria more than 1,000 patients were affected, in Slovakia tens of human cases occurred, so the probability of developing the chronic form of Q fever in Slovakia was lower. The small number of patients in Slovakia could also be explained by earlier diagnosis and proper antibiotic treatment at the early stage of infection. The patients' history, e.g., previous rheumatic disease, should be also taken into consideration. Lessons Learned from Q-Fever Outbreaks in Bulgaria and Slovakia Epidemiologic and serologic investigations in Bulgaria and Slovakia indicate that an increase in human Q fever in Bulgaria in the 1990s and Slovakia in 1993 was associated with goats. The data on the propensity of goats to transmit C. burnetii to humans from Greece, Cyprus, France, the United States, and even a trans-Pacific cargo ship transporting dairy goats, were summarized by Lang (4). More recently, a cluster of human C. burnetii infections associated with exposure to vaccinated goats and their unpasteurized products was reported from France (30). Goats may pose a threat to human health as a source of C. burnetii infection in every country in which they are raised extensively and are in close contact with humans. However, Q fever can also be contracted from other sources of infection and has been, even in Bulgaria and Slovakia. Reporting of Q fever in a given territory depends on the attention of public health authorities and the availability of diagnostic methods. Apart from C. burnetii isolation (mainly in cell cultures by a shell-vial method [31] or direct detection, preferably by polymerase chain reaction [32]), these diagnostic methods are based mostly on serologic tests. Sensitivity of serologic tests for screening Q-fever antibodies increased from CF to MA and from MIF to ELISA (27). The cut-off values for individual tests may differ between laboratories and antigens used; for CF and MA tests, 1:8 and 1:16 serum dilutions were acceptable (in Slovakia) and for either test 1:10 serum dilution was acceptable (in Bulgaria). For a more sensitive MIF and ELISA allowing also detection of immunoglobulin classes, diagnostic titers were set at the phase-II IgG >/= 200 and phase-II IgM >/= 50 in MIF (33) and at >/= 128 for the IgM and IgG phase-I responses, but >/= 512 for the IgM and >/= 1,024 for the IgG response to phase II C. burnetii in ELISA (34), respectively. Serologic diagnosis of acute Q fever relies on seroconversion from negativity to positivity or at least fourfold rise of phase-II antibodies in paired (acute- and convalescent-phase) serum samples and demonstration of IgM antibody response or high titers (e.g., >/= 128 in CF and 200 in MIF) of phase-II antibodies in a single serum sample. For diagnosis of chronic Q fever, high titers (i.e., >/= 200 in CF and >/= 800 of in MIF) of phase I antibodies, occurring rarely and in low titers in acute Q-fever cases, are required (9). Q-fever control and prevention measures have been reviewed (35). Apart from the thorough control of imported domestic animals, raw materials, and movement of domestic animals within a country, prevention measures should include adequate disinfection and disposal of animal products of conception and strict hygienic measures in cattle, sheep, and goat farms; plants processing products of these animals; boiling or pasteurization of milk at 62.8°C for 30 minutes or at 71.7°C for 15 seconds; and vaccination. At present, three types of Q-fever vaccine are available for human use: a Formalin-inactivated whole-cell phase-I C. burnetii vaccine used in Australia (36), a chloroform-methanol residue subunit of phase-I C. burnetii recommended by American authors (37), and Q-fever chemovaccine (a soluble subunit vaccine obtained by treatment with trichloroacetic acid of phase-I cells) developed and used in Romania (38) and the former Czechoslovakia (39). For vaccination of domestic animals, corpuscular phase I (in Slovakia) or phase II (e.g., in France) were used. The fact that phase-II vaccine did not protect goats from shedding C. burnetii in milk (30) confirmed that an effective Q-fever vaccine should consist of or be prepared from phase-I C. burnetii (40). Efficient recombinant vaccines, however, should also be pursued. Mass vaccination of cattle in Slovakia in the 1970s, followed by selective vaccination of cattle in serologically positive herds and elimination of positive reactors in the 1980s could lessen not only distribution of C. burnetii among domestic animals, but also its transmission to humans. However, absence of vaccination of domestic animals in Bulgaria could contribute to the maintenance of C. burnetii and therefore to increased possibility of human infection, though basic natural conditions for circulation of this agent in either country have been similar. Moreover, gradual changes in agriculture in Slovakia during the 1990s resulted in reduced numbers of cattle and sheep but not in the dramatic increase in goat numbers seen in Bulgaria after the collapse of state farms and cooperative units. One can conclude that in Bulgaria there is a permanent threat of more Q-fever outbreaks unless preventive measures, including improvement of veterinary services and vaccination of domestic animals, particularly goats, are established. In Slovakia, because of surveillance, veterinary control, and vaccination of domestic animals, the situation is much better; however, attention should still be paid to avoid introduction of C. burnetii by imported animals and raw materials and the possibility of coxiellosis outbreaks among domestic animals and consequently Q fever in humans. --------------------------------------------------------------------------- Dr. Serbezov is consultant at the National Center of Infectious and Parasitic Diseases in Sofia. Address for correspondence: Ján Kazár, Institute of Preventive and Clinical Medicine, Limbová 14, 833 01 Bratislava, Slovak Republic; fax: 421-737-3906; e-mail: kazar@upkm.sanet.sk. References 1. Derrick EH. Q fever, a new fever entity: clinical features, diagnosis and laboratory investigation. Med J Aust 1937;2:281-99. 2. Marrie TJ. Epidemiology of Q fever. In: Marrie TJ, editor. Q Fever. Vol 1. The disease. Boca Raton (FL): CRC Press; 1990. p. 49-70. 3. Hilbink F, Penrose M, Kovácová E, Kazár J. Q fever is absent from New Zealand. Int J Epidemiol 1993;22:945-9. 4. Lang GH. Coxiellosis (Q fever) in animals. In: Marrie TJ, editor. Vol 1. The disease. Boca Raton (FL): CRC Press; 1990. p. 23-48. 5. Sawyer LA, Fishbein DB, McDade JE. Q fever: current concepts. Review of Infectious Diseases 1987;9:935-46. 6. Rehácek J, Tarasevich IV. Q fever. In: Grešíková M, editor. Acari-borne rickettsiae & rickettsioses in Eurasia. Bratislava: VEDA; 1988. p.203-43. 7. Raoult D. Q fever: still a query after all those years. J Med Microbiol 1996;44:77-8. 8. Dupuis G, Petite J, Peter O, Vouilloz M. An important outbreak of human Q fever in a Swiss Alpine Valley. Int J Epidemiol 1987;16:282-9. 9. Raoult D, Marrie T. Q fever. Clin Infect Dis 1995;20:489-96. 10. Raoult D. Treatment of Q fever. Antimicrob Agents Chemother 1993;37:1733-6. 11. Kazár J. Q fever. In: Kazár J, Toman R, editors. Rickettsiae and rickettsial diseases. Bratislava: VEDA; 1996. p. 353-62. 12. Imhauser K. Viruspneumonien: Q-Fieber und Virusgrippe. Klinische Wochenschrift 1949;27:353-60. 13. Bieling R. Die Balkangrippe das Q Fieber der alten Welt. Beitrage für Hygiene und Epidemiologie 1950,H5. 14. Robbins FC, Ragan CA. Q fever in the Mediterranean area: report of its occurrence in allied troops. I. Clinical features of the disease. American Journal of Hygiene 1946;44:6-22. 15. Robbins FC, Gauld RL, Warner FB. Q fever in the Mediterranean area. II. Epidemiology. American Journal of Hygiene 1946;44:23-50. 16. Combiescu D, Vasiliu V, Dumitrescu N. Identification l'une nouvelle rickettsiose chez homme en Roumanie. Comptes Rendus des séances de la Sociéte de biologie Bucharest 1947;141:716-7. 17. Caminopetros JP. La Q-fever en Grece: le lait source de l'infection pour l'homme et les animaux. Annals du Parasitologie Paris 1948;23:107-18. 18. Mitov A. Diagnosis of two cases of Q fever in southern Bulgaria. Bulgarskaja Klinika (in Bulgarian) 1949;8:610-23. 19. Serbezov V, Shishmanov E, Aleksandrov E, Novkirishki V. Rickettsioses in Bulgaria and other Balkan countries. Danov CG, editor. Plovdiv: Christo G. Domov; 1973. p. 223 (in Bulgarian). 20. Georgieva G. Ixodid ticks as vectors of rickettsiae in Bulgaria [thesis in Bulgarian]; Sofia, Bulgaria: Military Medical Institute; 1984. 21. Rehácek J, Brezina R, Kovacova E, Zupanicicova M. Haemocyte test—an easy, quick and reliable method for the detection of rickettsiae in ticks. Acta Virol 1971;15:237-40. 22. Novkirishki V, Bojadzhian CH, Kijanovska E, et al. Epidemiologic studies of Q fever outbreak in the region of the Knyezha town. Infektologija (in Bulgarian) 1994;31:16-9. 23. Serbezov V. Q-fever endocarditis: etiology, epidemiology and etiological diagnostics. Bulgarian Cardiology (in Bulgarian) 1996;3:36-41. 24. Stein A, Raoult D. Q fever and pregnancy in humans and animals. In: Kazár J, Toman R, editors. Rickettsiae and rickettsial diseases. Bratislava: VEDA; 1996. p. 551-7. 25. Rehácek J. Epidemiology and significance of Q fever in Czechoslovakia. Zentralblatt für Bakteriologie, Mikrobiologie und Hygiene Series A 1987;267:16-9. 26. Rehácek J, Úrvolgyi J, Kocianová E, Sekeyová Z, Vavreková M, Kovácová E. Extensive examination of different tick species for infestation with Coxiella burnetii in Slovakia. Eur J Epidemiol 1991;7:299-303. 27. Kovácová E, Kazár J, Španìlová D. Analysis of antibody response in humans and goats with the use of different Coxiella burnetii antigenic preparations. In: Kazár J, Toman R, editors. Rickettsiae and rickettsial diseases. Bratislava: VEDA; 1996. p. 463-8. 28. Varga V. An explosive outbreak of Q fever in Jedlóvé Kostol'any, Slovakia. Cent Eur J Publ Health 1997;3:180-2. 29. Kazár J, Hornícek J, Valihrach J, Krunert Z, Pavlík J, Petrík P, et al. An epidemic of Q fever in a cotton-processing plant. Ceskoslovenská Epidemiologie, Mikrobiologie a Imunologie (in Slovak) 1982;31:144-51. 30. Fishbein DB, Raoult D. A cluster of Coxiella burnetii infections associated with exposure to vaccinated goats and their unpasteurized products. Am J Trop Med Hyg 1992;47:35-40. 31. Musso D, Raoult D. Coxiella burnetii cultures from acute and chronic Q-fever patients. J Clin Microbiol 1995;33:3129-32. 32. Stein A, Raoult D. Detection of Coxiella burnetii by DNA amplification using polymerase chain reaction. J Clin Microbiol 1992;30:2462-6. 33. Tissot Dupont H, Thirion X, Raoult D. Q fever serology: cutoff determination for microimmuno-fluorescence. Clin Diagn Lab Immunol 1994;1:89-96. 34. Waag D, Chulay J, Marrie T, England M, Williams J. Validation of an enzyme immunoassay for serodiagnosis of acute Q fever. Eur J Clin Microbiol Infect Dis 1995;14:421-7. 35. Kazár J, Brezina R. Control of rickettsial diseases. Eur J Epidemiol 1991;7:282-6. 36. Marmion BP, Ormsbee RA, Kyrkou M, Wright J, Worswick DA, Izzo AA, et al. Vaccine prophylaxis of abattoir-associated Q fever: eight years of experience in Australian abattoirs. Epidemiol Infect 1990;104:275-87. 37. Fries LF, Waag DM, Williams JC. Safety and immunogenicity in human volunteers of a chloroform-methanol residue vaccine for Q fever. Infect Immun 1993;61:1251-8. 38. Cracea E, Dumitrescu S, Botez D, Toma E, Bandu C, Sabin S, et al. Immunization in man with a soluble Q fever vaccine. Archives Roumaines de Pathologie Experimentale et de Microbiologie 1973;32:45-51. 39. Brezina R, Schramek Š, Kazár J, Úrvolgyi J. Q fever chemovaccine for human use. Acta Virol 1974;18:26. 40. Kazár J, Rehácek J. Q fever vaccines: present status and application in man. Zentralblatt fur Bakteriologie, Mikrobiologie und Hygiene Series A 1987;267:74-8. --------------------------------------------------------------------------- Synopses Adhesins as Targets for Vaccine Development Theresa M. Wizemann, John E. Adamou, and Solomon Langermann MedImmune, Inc., Gaithersburg, Maryland, USA --------------------------------------------------------------------------- Blocking the primary stages of infection, namely bacterial attachment to host cell receptors and colonization of the mucosal surface, may be the most effective strategy to prevent bacterial infections. Bacterial attachment usually involves an interaction between a bacterial surface protein called an adhesin and the host cell receptor. Recent preclinical vaccine studies with the FimH adhesin (derived from uropathogenic Escherichia coli) have confirmed that antibodies elicited against an adhesin can impede colonization, block infection, and prevent disease. The studies indicate that prophylactic vaccination with adhesins can block bacterial infections. With recent advances in the identification, characterization, and isolation of other adhesins, similar approaches are being explored to prevent infections, from otitis media and dental caries to pneumonia and sepsis. The ultimate aim of any vaccine is to produce long-term protective immune responses against a pathogen. These responses include systemic humoral antibodies that neutralize invasive organisms and cytotoxic T cells, which are required to clear certain infections, particularly chronic viral infections such as HIV (1). Helper T-cell and cytokine responses also influence humoral and cellular immune responses (2). For most bacteria and viruses, the first encounter with their host involves attachment to a eukaryotic cell surface, which results in colonization of the host prior to disease. In such cases, induced antibody responses at the mucosal surface could prevent attachment and abrogate colonization. The ideal target for such antibodies—surface proteins known as adhesins—mediate microbial attachment to host tissue (3). We review recent advances in the identification, isolation, and purification of adhesins (and putative adhesins) that could serve as vaccines to elicit such responses. Studies of at least one adhesin, the FimH protein from uropathogenic Escherichia coli, show that anti–adhesin antibodies can block microbial attachment and subsequent disease. Furthermore, while specific induction of immune responses along the mucosal surface by mucosal immunization may have its advantages (4-7), the FimH studies demonstrate that immunoglobulin (Ig) G antibodies alone, which transudate into secretions after parenteral vaccination with the FimH adhesin, are sufficient to block colonization and infection. The Role of Adhesins in Microbial Pathogenesis To initiate infection, bacterial pathogens must first be able to colonize an appropriate target tissue of the host (8,9). This tropism (ability to gain access to a niche within the body), in association with the ability of the bacterium to breach mucosal barriers and invade the host, distinguishes pathogenic from commensal organisms. Colonization begins with the attachment of the bacterium to receptors expressed by cells forming the lining of the mucosa (Figure 1A, 1B). Certain species of bacteria are restricted in terms of the hosts and tissues they infect and the diseases they cause. In many cases, tropism for specific tissues has been corroborated in the laboratory by in vitro binding assays with isolated epithelial cells collected from sites of infection or from infection-prone hosts. Attachment is mediated by adhesin proteins; bacterial lectins are the most common type of adhesin among both gram-negative and gram-positive bacteria (3,10-12). Adhesins, such as the FimH adhesin produced by most Enterobacteriaceae (including uropathogenic E. coli), are highly conserved proteins (13). This lack of major variation is most likely due to the requirement that all pathogenic strains recognize invariant host receptors. Although minor changes in the adhesin protein have been observed (2% divergence) and correlate with decreased or increased affinity for binding to sugars (14), antibodies against a single FimH protein cross-react with >90% of E. coli strains expressing the FimH adhesin and block binding to bladder cells in vitro (15,16). Furthermore, antibodies against FimH from a single isolate protect against in vivo colonization by >90% of uropathogenic strains in a murine model for cystitis (unpub. obs.). This high degree of antigenic conservation is another reason why adhesins may serve as ideal vaccines. Figure 1. Four mechanisms of bacterial adherence where anti-adhesin vaccines could potentially block colonization and infection. A shows pili or fibrillae protruding from the bacterial surface. These proteinaceous appendages bind to host cell surface molecules, usually carbohydrates, by adhesin proteins located at the distal tip of the pilus/fibrillar organelle. Antibodies targeting the adhesin protein block the bacterial/host interaction. B demonstrates a similar process of bacterial/epithelial cell interactions mediated by afimbrial adhesin proteins. In this case, antibodies directed against the bacterial surface proteins should also block attachment and colonization by impeding the ability of the bacteria to associate with mucosal tissues. C illustrates that some bacteria establish intimate associations with eukaryotic cells byintimin proteins, resulting in cytoskeletal rearrangements, host cell signaling, possible internalization of the bacteria, and in many cases systemic disease. Blocking the intimate association/adherence may also be another strategy to prevent bacterial infections. D shows a novel mechanism whereby bacteria secrete their own receptor protein, which is internalized by the target host cell, phosphorylated, and embedded in the eukaryotic cell as a new receptor for tight binding by the bacterium. Theoretically, blocking the secreted receptor (Hp90) before it is internalized by the host cell could provide another mechanism to block bacterial adherence and infection. Aside from mediating colonization of the host, bacterial attachment often results in the up-regulation or expression of many other virulence genes encoding various proteins that allow for invasion of the host (Figure 1C) (17-23). The proteins can mediate tighter associations with epithelial cells, trigger epithelial cell–actin filament rearrangements, and induce changes in host-cell signaling and function. In some cases, the bacteria may also secrete a protein that inserts into mammalian cells and serves as a receptor for its own intimate adherence with the host (Figure 1D) (24). Given that cross-talk between pathogenic bacteria and host cells after microbial attachment may trigger expression of virulence factors leading to local inflammation or invasive disease, vaccines that block bacterial attachment may have multiple advantages. Many studies have demonstrated the utility of vaccines against bacterial surface components in blocking attachment in vitro as well as in vivo (25-28). However, as understanding of the mechanisms of attachment has evolved and characterization of specific adhesin molecules has been refined, new opportunities have emerged for the development of adhesin-based vaccines. Bacterial Adhesins: Gram-Negative Organisms One of the best understood mechanisms of bacterial adherence is attachment mediated by cell surface structures called pili or fimbriae. Pili are long, flexible structures that extend outward from the bacterial surface of many species of bacteria and allow for contact between the bacteria and the host cell. Originally pili were thought to be homopolymeric structures composed of approximately 1,000 copies of a single structural subunit (fimbrin) packed in a helical array. However, for many pili, such as the highly characterized Type 1 and Pap pili expressed on E. coli and other Enterobacteriaceae, they are heteropolymeric structures with minor tip fibrillae proteins located at the distal end of the organelle (11). The specific interaction with receptor architectures on host cell surfaces is mediated by one of these tip proteins, called adhesins. For example, FimH adhesin mediates attachment to alpha-D-mannosides (29,30) by type 1 pili, while PapG mediates binding to alpha-D-gal(1-4)ß-D-Gal-containing receptors on host cells by Pap pili (31,32) (Table 1). The specificity of interaction may be involved in conferring tropism to the bladder and kidney tissues, respectively (33,38). Pilus-associated adhesins have been identified in a number of other bacteria as well (Table 1). Not all adhesins are associated with pili. Bordetella pertussis expresses at least two putative adhesins on its surface: filamentous hemagglutinin (FHA) and pertactin (34). FHA is thought to mediate attachment to sulfated sugars on cell-surface glycoconjugates, although it may also have other properties. Pertactin is thought to mediate binding by the Arg-Gly-Asp triplet binding sequence characteristic of integrin-binding proteins, although the role of this binding activity in the pathogenesis of Bordetella infections is unclear. Both FHA and pertactin are components in the recently approved acellular pertussis vaccine. Table 1. Adhesins of gram-negative bacteria -------------------------------------------------------------------------- Adhesin Strain Ligand Reference -------------------------------------------------------------------------- Pili family(sup a) Hultgren (33) PapG Escherichia Gala(1-4)Gal in globoseries of coli glycolipids SfaS E. coli a-sialyl-2 3-b-galactose FimH E. coli Mannose-oligosaccharides HifE Haemophilus Sialylyganglioside-GM1 influenzae PrsG E. coli Gala(1-4)Gal in globoseries of glycolipids MrkD Klebsiella Type V collagen pneumoniae FHA Bordetella Sulfated sugars on Brennan pertussis cell-surface glycoconjugates (34) Pertactin B. pertussis Integrins Brennan (34) HMW1/HMW2 H. influenzae Human epithelial cells St. Geme (35) Hia H. influenzae Human conjunctival cells Barenkamp (36) Le(sup b)- Helicobacter Fucosylated Le(sup b) histo-blood binding Ilver (37) adhesin pylori group antigens -------------------------------------------------------------------------- (sup a)Representative examples from the large family of pilus-associated adhesins. FHA, filamentous hemagglutinin; HMW, high molecular weight; Hia, H. influenzae adhesin In Haemophilus influenzae, two families of nonpilus adhesins have been identified: high-molecular weight adhesion proteins (HMW1 and HMW2) and immunogenic high molecular-weight surface-exposed proteins, the prototypic member of which has been designated Hia for H. influenzae adhesin (35). Both families are expressed by nontypable H. influenzae, which colonize the respiratory tract and cause such diseases as otitis media, pneumonia, and bronchitis. The HMW proteins, which share homology with the B. pertussis FHA protein, mediate specific attachment of H. influenzae to different types of human epithelial cells in vitro and have been implicated in directing respiratory tract tropism for these organisms. The Hia protein, in contrast, mediates tight association with human conjunctival cells and is present only in H. influenzae strains deficient in HMW1/HMW2 expression (36). Given the dichotomy among nontypable strains expressing either HMW or Hia-like adhesins and the serologic conservation at least among the HMW1 and HMW2 proteins, a vaccine based on a combination of such proteins may be protective against disease caused by most nontypable H. influenzae. Another nonpilus adhesin has been identified and purified from Helicobacter pylori. The adhesin, called Le(sup b)-binding adhesin, mediates bacterial adherence to fucosylated Lewis b (Le(sup b)) histoblood group antigens, which are expressed along the mucosal surface of the gastric epithelium (37). The Leb-binding adhesin may be involved in conferring tropism for stomach epithelium and allowing pathogenic bacteria to establish an ecologic niche within the gastrointestinal tract. In association with other virulence determinants expressed by H. pylori in the stomach, the colonization process ultimately results in ulcer formation. A unique mechanism has been identified by which certain strains of enteropathogenic E. coli that cause severe diarrhea target cells for attachment: Enteropathogenic E. coli express and insert their own receptor (Hp90) into mammalian cell surfaces, thereby allowing the bacteria to attach and establish intimate contact with the epithelial cells (24) (Figure 1D). Although a bacterially encoded receptor for a cognate adhesin protein (intimin), Hp90 is expressed and secreted by enteropathogenic E. coli before colonization; therefore, it may also serve as a target for vaccine development. Bacterial Adhesins: Gram-Positive Organisms Some of the most well-characterized colonization factors in gram-positive bacteria—the polypeptides of the antigen I/II family—bind to salivary glycoproteins in a lectinlike interaction (11) and promote adhesion to the tooth surface (Table 2). These proteins include the original AgI/II from Streptococcus mutans, also known as SpaP, P1, or PAc, and the Streptococcus sobrinus SpaA and PAg proteins (39). Streptococcus gordonii expresses two antigen I/II polypeptides, SspA and SspB, products of tandem chromosomal genes (39). Surface proteins of the antigen I/II family contain alanine-rich repeats, which adopt an a-helical coiled-coil structure, proline rich repeats, and a carboxy-terminal region that includes the gram-positive cell wall anchor motif LPXTG (12,49). Binding activity to salivary glycoprotein has been attributed to both the highly conserved alanine rich repeats (50,51) and the proline-rich repeating sequences (52). In addition to salivary glycoprotein binding activity, SspA has been implicated in coaggregation of S. gordonii with Actinomyces (39). Such bacterial coaggregation may be involved in dental plaque formation. Two other S. gordonii proteins, CshA and CshB, have also been implicated in coaggregation with Actinomyces. These adhesins may play a role in adherence of S. gordonii to immobilized fibronectin in vitro (52) and in colonization in vivo (40). Staphylococcus aureus also expresses fibronectin-binding adhesins. Two genes encoding for fibronectin-binding proteins have been identified in S. aureus fnbA and fnbB (41). Fibronectin binding activity is critical in pathogenesis because it allows the bacteria to adhere to extracellular matrix components including fibronectin and collagen (53). This can result in cutaneous infections and in life-threatening bacteremia and endocarditis (54). Table 2: Adhesins of gram-positive bacteria -------------------------------------------------------------------------- Adhesin Strain Ligand Reference -------------------------------------------------------------------------- Antigen I/II Family Demuth (39) Ag I/II, SpaP, P1, PAc Streptococcus Salivary mutans glycoprotein Salivary SspA, SspB Streptococcus glycoprotein gordonii Actinomyces SpaA, PAg Streptococcus Salivary sobrinus glycoprotein CshA , CshB S. gordonii Actinomyces, McNab (40) fibronectin FnbA, FnbB Staphylococcus Fibronectin Jonsson (41) aureus SfbI, Protein F Streptococcus Fibronectin Talay (42) Hanski pyogenes (43) LraI gamily Burnett-Curley et al. (44) Jenkinson and Lamont (12) Salivary FimA Streptococcus glycoprotein parasanguis fibrin PsaA Streptococcus Unknown pneumoniae ScaA S. gordonii Actinomyces SsaB Streptococcus Salivary sanguis glycoprotein EfaA Enterococcus Unknown faecalis CbpA/SpsA/PbcA/ S. pneumoniae Cytokine activated Rosenow (45) PspC Hammerschmidt et epithelial and al. (46) endothelial Briles et al. (47) cells, IgA Smith et al. (48) -------------------------------------------------------------------------- Binding to fibronectin is also essential for the attachment of S. pyogenes to respiratory endothelial cells. S. pyogenes, a group A streptococcus, has been implicated in various diseases from skin and throat infections to sepsis and shock (55). This binding activity is mediated by several fibronectin binding proteins. Six different genes encoding proteins with fibronectin binding activity have been identified (12). The most well characterized are two closely related proteins, SfbI (42) and Protein F (43). Both proteins are directly involved in the fibronectin-mediated adherence to epithelial cells. Another group of extensively studied streptococcal adhesins is the LraI family of proteins. Included in this group are FimA from S. parasanguis, PsaA from S. pneumoniae, ScaA from S. gordonii, ScbA from S. crista, SsaB from S. sanguis, and EfaA from Enterococcus faecalis (44,56). These membrane-bound lipoproteins are part of a larger family of ATP-binding cassette metal permeases, involved in the acquisition of manganese (57). The adhesins FimA and SsaB have high affinity for salivary glycoprotein on tooth surfaces and are involved in colonization of the oral cavity (58). The FimA adhesin for S. parasanguis has been localized to the tip of peritrichous surface fimbria on these bacteria (58). FimA is also a major virulence factor of S. parasanguis and binds fibrin (44). The ability to bind fibrin has been implicated in the pathogenesis of infective endocarditis. The S. pneumoniae homologue of FimA, PsaA, also has a role in pathogenesis (59,60). While it commonly colonizes the nasopharynx of healthy persons, S. pneumoniae is a common pathogen in children and older adults and a leading cause of otitis media, bacterial pneumonia, sepsis, and meningitis. Like its homologues, PsaA may function as an adhesin, according to initial evidence (61). Insertion inactivation of psaA resulted in pneumococcal mutants that exhibited reduced adherence to alveolar epithelial cells in vitro (60). However, the PsaA protein may be a permease involved in the regulation of adherence rather than functioning as an adhesin per se (57,61). Another pneumococcal protein thought to be an adhesin is CbpA (45) (also known as SpsA [46] and PspC [47]), one of a family of choline binding proteins (CBPs) (i.e., surface proteins noncovalently associated with the phosphocholine on the lipoteichoic acid). CbpA was initially isolated from a mixture of pneumococcal proteins that were able to bind to a choline affinity column. The CBP mix was purified from a strain with an inactivated pspA gene (45). The exogenous CBP mix inhibited adherence of pneumococci to type II pneumocytes and endothelial cells in vitro, suggesting that one or more of these proteins may act as adhesins. CbpA was the most abundant component in this mix and was shown to be on the surface of intact pneumococci in an indirect fluorescent-labeling assay. The CbpA protein also reacted strongly with a pool of human convalescent-phase serum. CbpA is thought to mediate adherence of S. pneumoniae to sialic acid and lacto-N-neotetraose ligands present on cytokine-activated epithelial and endothelial cells in vitro (45). cbpA-defective mutants did not colonize the nasopharynx of infant rats, further supporting its function as an adhesin and potential usefulness as an adhesin-based vaccine (45). In addition, others have shown that this choline binding protein (which they called SpsA) has IgA-binding properties, though the relevance of this function to pneumococcal pathogenesis is unclear (46). Yet another group has demonstrated complement protein C3-binding activity for this protein, which they termed PbcA (48). Whatever its role in pathogenesis, we have demonstrated that the gene for CbpA (SpsA/PbcA/PspC) is highly conserved among the most common pneumococcal isolates, further enhancing its use as a vaccine candidate (62). Adhesins as Vaccines: FimH as a Paradigm for Adhesin-Based Vaccines To Block Colonization One of the key aspects of proving the potential efficacy of an adhesin-based vaccine in vivo is the development of an animal model of disease that relies on bacterial colonization of the mucosal epithelium mediated by the specific adhesin of interest. Although seemingly straightforward, testing for protection in small animal models of disease is difficult for various reasons: large doses of in vitro grown bacteria are required to establish mucosal colonization in animals, which does not necessarily mimic the course of infection in humans; specific glycoprotein receptors for some adhesins are lacking in animal mucosal tissues that correspond to the site of colonization in humans; and some bacterial adhesins that are usually expressed as part of a larger structure on the bacterial cell surface (e.g., tip adhesins associated with whole pili) are difficult to purify. Despite these difficulties, adhesin-based vaccines have demonstrated efficacy in protecting against infection, thus proving the usefulness of such molecules as subunit vaccines. Research using the FimH adhesin from E. coli provides one such example. Type 1 pili have long been implicated in bacterial urinary tract infections in humans (63,64). In a murine cystitis model, colonization of the bladder by E. coli was shown to depend on growth conditions that favored expression of type 1 pili and in particular required FimH (15,65). Thus, the murine model was a valid small-animal model to prove whether adhesin-based vaccines might block colonization. Although purifying large amounts of pilus-associated adhesin is difficult (because most adhesins are proteolytically degraded when expressed as independent moieties), Hultgren et al. demonstrated that the FimH adhesin could be stabilized in an active conformation by the periplasmic chaperone FimC, making it possible to purify full-length FimH protein. Vaccination with the FimCH complex elicited long-lasting immune responses to FimH. Sera from mice vaccinated with the FimH vaccine inhibited uropathogenic strains of E. coli from binding to human bladder cells in vitro. Vaccination with the FimH adhesin-vaccine reduced in vivo colonization of the bladder mucosa by >99% in the murine cystitis model (15). Furthermore, the FimH vaccine protected against colonization and disease by uropathogenic strains of E. coli capable of expressing multiple adhesins. IgG specific for FimH was detected in the urine of protected mice, consistent with our original hypothesis that antibodies directed against an adhesin protein might protect along the mucosal surface. Subsequent studies in a primate model of cystitis have corroborated these findings (Langermann et al., unpub. data). Furthermore, in primate studies we demonstrated a direct correlation between the presence of inhibitory antibodies in secretions and protection against colonization and infection. While IgG antibodies elicited against adhesins are protective, induction of immune responses along the mucosa can be augmented by a variety of antigen delivery systems that specifically target mucosa-associated lymphoid tissue and activate the mucosal immune system (4-7,66). These delivery systems include whole-inactivated or live-attenuated bacterial and viral vectors, biodegradable microspheres, liposomes, transgenic plants, and antigens conjugated to or coadministered with the cholera toxin B subunit or attenuated forms of heat labile toxin from E. coli. Many of these systems hold promise for future vaccine strategies, but only a few have been tested in humans for safety and adjuvanticity. As these mucosal adjuvants progress further toward approval for use in humans, testing should be done with adhesin antigens to determine if induction of local immune responses enhances the protective efficacy of adhesin-based vaccines as compared with conventional parenteral vaccination. Such studies are under way with the FimH vaccine. Given the preclinical data with the FimH vaccine, similar efforts should be directed at developing adhesin-based vaccines for a wide range of pathogens. In this regard, additional efforts should also be focused on developing mucosal models of infection. The availability of such models should allow for appropriate screening of adhesin-based vaccines to prevent infections by streptococci, staphylococci, and other pathogens for which vaccine coverage is absent or inadequate. --------------------------------------------------------------------------- Acknowledgments We thank Luis Branco for his illustration (Figure 1) and Scott Koenig and Jennifer Bostic for reviewing the manuscript. Dr. Wizemann served as Streptococcus pneumoniae Vaccine Project Team Leader at MedImmune from 1996-98. Address for correspondence: Solomon Langermann, MedImmune, Inc., Department of Immunology and Molecular Genetics, 35 West Watkins Mill Road, Gaithersburg, MD 20878, USA; fax: 301-527-4200; e-mail: langermanns@medimmune.com. References 1. Hilleman MR. Paper presented at the International Symposium on Recombinant Vectors in Vaccine Development; 1993 May 23; Albany, New York. [Nucleic Acids Technologies Foundation, sponsored by IABS, FDA, USDA/APHIS, and NIAID/NIH]. 2. Alexander J, Fikes J, Hoffman S, Franke E, Sacci J, Appella E, Chisari, FV, et al. The optimization of helper T lymphocyte (HTL) function in vaccine development. Immunol Res 1998;18:79-92. 3. St Geme JW III. Bacterial adhesins: determinants of microbial colonization and pathogenicity. Adv Pediatr 1997;44:43-72. 4. McGhee J, Mestecky J, Dertzbaugh MT, Eldridge JH, Hirasawa M, Kiyono H. The mucosal immune system: from fundamental concepts to vaccine development. Vaccine 1992;10:75-88. 5. Langermann S. New approaches to mucosal immunization. Semin Gastrointest Dis 1996;7:12-8. 6. O'Hagan DT. Recent advances in vaccine adjuvants for systemic and mucosal administration. J Pharm Pharmacol 1998;50:1-10. 7. Mestecky J, Michalek SM, Moldoveanu Z, Russell MW. Routes of immunization and antigen delivery systems for optimal mucosal immune responses in humans. Behring Inst Mitt 1997;98:33-43. 8. Beachey EH. Bacterial adherence: adhesin-receptor interactions mediating the attachment of bacteria to mucosal surface. J Infect Dis 1981;143:325-45. 9. Beachey EH, Giampapa CS, Abraham SN. Bacterial adherence: adhesin receptor-mediated attachment of pathogenic bacteria to mucosal surfaces. American Review of Respiratory Diseases 1988;138:S45-8. 10. Ofek I, Sharon N. Adhesins as lectins: specificity and role in infection. Curr Top Microbiol Immunol 1990;151:91-113. 11. Hultgren SJ, Abraham S, Capron M, Falk P, St. Geme JW, Normark S. Pilus and non-pilus bacterial adhesins: assembly and function in cell recognition. Cell 1993;73:887-901. 12. Jenkinson HF, Lamont RJ. Streptococcal adhesion and colonization. Crit Rev Oral Biol Med 1997;8:175-200. 13. Abraham SN, Sun D, Dale JB, Beachey EH. Conservation of the D-mannose-adhesion protein among type 1 fimbriated members of the family Enterobacteriacaea. Nature 1988;682-4. 14. Sokurenko EV, Courtney HS, Ohman DE, Klemm P, Hasty DL. FimH family of type 1 fimbrial adhesins: functional heterogeneity due to minor sequence variations among fimH genes. J Bacteriol 1994;176:748-55. 15. Langermann S, Palaszynski S, Barnhart M, Auguste G, Pinkner JS, Burlein J, et al. Prevention of mucosal Escherichia coli infection by FimH-based systemic vaccination. Science 1997;276:607-11. 16. Palaszynski S, Pinkner J, Leath S, Barren P, Auguste CG, Burlein J, et al. Systemic immunization with conserved pilus-associated adhesins protects against mucosal infections. Dev Biol Stand 1998;92:117-22. 17. Finlay B, Falkow S. Common themes in microbial pathogenicity. Microbiological Review 1989;53:210-30. 18. Hoepelman AI, Tuomanen EI. Consequences of microbial attachment: directing host cell functions with adhesins. Infect Immun 1992;60:1729-33. 19. Svanborg C, Hedlund M, Connell H, Agace W, Duane, RD, Nilsson A, et al. Bacterial adherence and mucosal cytokine responses. Receptors and transmembrane signaling. Ann N Y Acad Sci 1996;797:177-90. 20. Falkow S. Invasion and intracellular sorting of bacteria: searching for bacterial genes expressed during host/pathogen interactions. J Clin Invest 1997;100:239-43. 21. Mecsas JJ, Strauss EJ. Molecular mechanisms of bacterial virulence: type III secretion and pathogenicity islands. Emerg Infect Dis 1996;2:270-88. 22. Zhang JP, Normark S. Induction and gene expression in Escherichia coli after pilus-mediated adherence. Science 1996;273:1234-6. 23. Donnenberg MS, Kaper JB, Finlay BB. Interactions between enteropathogenic Escherichia coli and host epithelial cells. Trends Microbiol 1997;5:109-14. 24. Kenny B, DeVinney R, Stein M, Reinscheid DJ, Frey EA, Finlay BB. Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells. Cell 1997;91:511-20. 25. Svanborg-Eden C, Marild S, Korhonen TK. Adhesion inhibition by antibodies. Scand J Infect Dis 1982;33:72-8. 26. O'Hanley P, Lark D, Falkow S, Schoolnik G. Molecular basis of Escherichia coli colonization of the upper urinary tract in Balb/c mice. Gal-Gal pili immunization prevents Escherichia coli pyelonephritis in the Balb/c mouse model of human pyelonephritis. J Clin Invest 1985;75:347-60. 27. Pecha B, Low D, O'Hanley P. Gal-Gal pili vaccines prevent pyelonephritis by piliated Escherichia coli in a murine model. Single component Gal-Gal pili vaccines prevent pyelonephritis by homologous and heterologous piliated E. coli strains. J Clin Invest 1989;83:2102-8. 28. Moon HW, Bunn TO. Vaccines for preventing enterotoxigenic Escherichia coli infections in farm animals. Vaccine 1993;11:213-20. 29. Maurer L, Orndorff P. Identification and characterization of genes determining receptor binding and pilus length of Escherichia coli type 1 pili. J Bacteriol 1987;169:640-5. 30. Hanson MS, Brinton CC Jr. Identification and characterization of the Escherichia coli type 1 pilus adhesin protein. Nature 1988;322:265-8. 31. Bock K, Breimer ME, Brignole A, Hansson GC, Karlsson KA, Larson,G, et al. Specificity of binding of a strain of uropathogenic Escherichia coli to Gal 1-4Gal-containing glycosphingolipids. J Biol Chem 1985;260:8545-51. 32. Stromberg N, Marklund BI, Lund B, Ilver D, Hamers A, Gaastra W, et al. Host-specificity of uropathogenic Escherichia coli depends on differences in binding specificity to Gal 1-4Gal-containing isoreceptors. EMBO J 1990;9:2001-10. 33. Hultgren SJ, Jones CH. Utility of the immunoglobulin-like fold of chaperones in shaping organelles of attachment in pathogenic bacteria. American Society for Microbiology News 1195;61:457-64. 34. Brennan MJ, Shahin RD. Pertussis antigens that abrogate bacterial adherence and elicit immunity. Am J Respir Crit Care Med 1996;154:S145-9. 35. St. Geme JW III. Progress towards a vaccine for nontypable Haemophilus influenzae. The Finnish Medical Society DUODECIM. Ann Med 1996;28:31-7. 36. Barenkamp SJ, St Geme JW III. Identification of a second family of high-molecular-weight adhesion proteins expressed by non-typable Haemophilus influenzae. Mol Microbiol 1996;19:1215-23. 37. Ilver D, Arnqvist A, Ogren J, Frick IM, Kersulyte D, Incecik ET, et al. Helicobacter pylori adhesin binding fucosylated histo-blood group antigens revealed by retagging. Science 1998;279:373-7. 38. Roberts JA. Tropism in bacterial infections: urinary tract infections. J Urol 1996;156:1552-9. 39. Demuth DR, Duan Y, Brooks W, Holmes AR, McNab R, Jenkinson HF. Tandem genes encode cell-surface polypeptides SspA and SspB which mediate adhesion of the oral bacterium Streptococcus gordonii to human and bacterial receptors. Mol Microbiol 1996;20:403-13. 40. McNab R, Jenkinson HF, Loach DM, Tannock GW. Cell-surface-associated polypeptides CshA and CshB of high molecular mass are colonization determinants in the oral bacterium Streptococcus gordonii. Mol Microbiol 1994;14:743-5. 41. Jonsson K, Signas C, Muller HP, Lindberg M. Two different genes encode fibronectin binding proteins in Staphylococcus aureus. The complete nucleotide sequence and characterization of the second gene. Eur J Biochem 1991;202:1041-8. 42. Talay SR, Valentin-Weigand P, Timmis KN, Chhatwal GS. Domain structure and conserved epitopes of Sfb protein, the fibronectin-binding protein adhesin of Streptococcus pyogenes. Mol Microbiol 1994;13:531-9. 43. Hanski E, Caparon M. Protein F, a fibronectin-binding protein, is an adhesin of the group A streptococcus Streptococcus pyogenes. Proc Natl Acad Sci U S A 1992;89:6172-6. 44. Burnette-Curley D, Wells V, Viscount H, Munro CL, Fenno JC, Fives-Taylor P, et al. FimA, a major virulence factor associated with Streptococcus parasanguis endocarditis. Infect Immun 1995;63:4669-74. 45. Rosenow C, Ryan P, Weiser JN, Johnson S, Fontan P, Ortqvist A, et al. Contribution of novel choline-binding proteins to adherence, colonization and immunogenicity of Streptococcus pneumoniae. Mol Microbiol 1997;25:819-29. 46. Hammerschmidt S, Talay SR, Brandtzaeg P, Chhatwal GS. SpsA, a novel pneumococcal surface protein with specific binding to secretory immunoglobulin A and secretory component. Mol Microbiol 1997;25:1113-24. 47. Briles DE, Hollingshead SK, Swiatlo E, Brooks-Walter A, Szalai A, Virolainen A, et al. PspA and PspC: their potential for use as pneumococcal vaccines. Microb Drug Resist 1997;3:401-8. 48. Smith BL, Cheng Q, Hostetter MK. Characterization of a pneumococcal surface protein that binds complement protein C3 and its role in adhesion [poster D-122]. The American Society for Microbiology (1998) 98th General Meeting; Atlanta, Georgia. 49. Hajishengallis G, Russell MW, Michalek SM. Comparison of an adherence domain and a structural region of Streptococcus mutans antigen I/II in protective immunity against dental caries in rats after intranasal immunization. Infect Immun 1998;66:1740-3. 50. Crowley PJ, Brady LJ, Piacentini DA, Bleiweis AS. Identification of a salivary agglutinin-binding domain within cell surface adhesin P1 of Streptococcus mutans. Infect Immun 1993;61:1547-52. 51. Todryk SM, Kelly CG, Lehner T. Effect of route of immunisation and adjuvant on T and B cell epitope recognition within a streptococcal antigen. Vaccine 1998;16:174-80. 52. McNab R, Holmes AR, Clark JM, Tannock GW, Jenkinson HF. Cell surface polypeptide CshA mediates binding of Streptococcus gordonii to other oral bacteria and to immobilized fibronectin. Infect Immun 1996;64:4202-10. 53. Schennings T, Heimdahl A, Coster K, Flock J-I. Immunization with fibronectin binding protein from Staphylococcus aureus protects against experimental endocarditis in rats. Microb Pathog 1993;15:227-36. 54. Lee JC. The prospects for developing a vaccine against Staphylococcus aureus. Trends Microbiol 1996;4:162-6. 55. Molinari G, Talay SR, Valentin-Weigand P, Rohde M, Chhatwal GS. The fibronectin-binding protein of Streptococcus pyogenes, SfbI, is involved in the internalization of group A streptococci by epithelial cells. Infect Immun 1997;65:1357-63. 56. Correia FF, DiRienzo JM, McKay TL, Rosan B. scbA from Streptococcus crista CC5A: an atypical member of the lraI gene family. Infect Immun 1996;64:2114-21. 57. Dintilhac A, Alloing G, Granadel C, Claverys JP. Competence and virulence of Streptococcus pneumoniae: Adc and PsaA mutants exhibit a requirement for Zn and Mn resulting from inactivation of putative ABC metal permeases. Mol Microbiol 1997;25:727-39. 58. Fenno JC, Shaikh A, Spatafora G, Fives-Taylor P. The fimA locus of Streptococcus parasanguis encodes an ATP-binding membrane transport system. Mol Microbiol 1995;15:849-63. 59. Talkington DF, Brown BG, Tharpe JA, Koenig A, Russell H. Protection of mice against fatal pneumococcal challenge by immunization with pneumococcal surface adhesin A (PsaA). Microb Pathog 1996;21:17-22. 60. Berry AM, Paton JC. Sequence heterogenity of PsaA, a 37-kilodalton putative adhesin essential for virulence of Streptococcus pneumoniae. Infect Immun 1996;64:5255-62. 61. Novak R, Brown JS, Charpentier E, Tuamonen E. Penicillin tolerance genes of Streptococcus pneumoniae: the ABC-type manganese permease complex Psa. Mol Microbiol 1998;29:1285-96. 62. Dormizter M, Wizemann TM, Adamou JE, Walsh W, Gayle T, Langermann S, et al. Sequence and structural analysis of CbpA, a novel choline-binding protein of Streptococcus pneumoniae [poster B-3]. The American Society for Microbiology (1998) 98th General Meeting; Atlanta, Georgia. 63. Schaeffer AJ, Amundsen SK, Schnidt LN. Adherence of Escherichia coli to human urinary tract epithelial cells. Infect Immun 1979;24:753-9. 64. Ofek I, Mosek A, Sharon N. Mannose-specific adherence of Escherichia coli freshly excreted in the urine of patients with urinary tract infections, and of isolates subcultured from the infected urine. Infect Immun 1981;34:708-11. 65. Connell H, Agace W, Klemm P, Schembri M, Marild S, Svanborg C. Type 1 fimbrial expression enhances Escherichia coli virulence for the urinary tract. Proc Natl Acad Sci U S A 1996;93:9827-32. 66. Frey A, Neutra MR. Targeting of mucosal vaccines to Peyer's patch M cells. Behring Inst Mitt 1997;98:376-89. _________________________________________________________________________ Research _________________________________________________________________________ Research Tuberculosis in the Caribbean: Using Spacer Oligonucleotide Typing to Understand Strain Origin and Transmission Christophe Sola, Anne Devallois, Lionel Horgen, Jérôme Maïsetti, Ingrid Filliol, Eric Legrand, and Nalin Rastogi Institut Pasteur de Guadeloupe, Pointe à Pitre, Guadeloupe --------------------------------------------------------------------------- We used direct repeat (DR)-based spacer oligonucleotide typing (spoligotyping) (in association with double-repetitive element–polymerase chain reaction, IS6110-restriction fragment length polymorphism [RFLP], and sometimes DR-RFLP and polymorphic GC-rich sequence-RFLP) to detect epidemiologic links and transmission patterns of Mycobacterium tuberculosis on Martinique, Guadeloupe, and French Guiana. In more than a third of the 218 strains we typed from this region, clusters and isolates shared genetic identity, which suggests epidemiologic links. However, because of limited epidemiologic information, only 14.2% of the strains could be directly linked. When spoligotyping patterns shared by two or more isolates were pooled with 392 spoligotypes from other parts of the world, new matches were detected, which suggests imported transmission. Persisting foci of endemic disease and increased active transmission due to high population flux and HIV-coinfection may be linked to the recent reemergence of tuberculosis in the Caribbean. We also found that several distinct families of spoligotypes are overrepresented in this region. Sequencing of the Mycobacterium tuberculosis H37Rv genome (1) has facilitated the study of the biodiversity of M. tuberculosis around the world (2). We investigated M. tuberculosis epidemiology and biodiversity in the Caribbean region, where sequencing data are not available. Caribbean islands possess independent and shared characteristics that justify investigation of both the molecular epidemiology (3) and the evolutionary history of M. tuberculosis (4). We used spacer oligonucleotide typing (spoligotyping), double-repetitiveelement (DRE)– polymerase chain reaction (PCR) and IS6110-restriction fragment length polymorphism (RFLP) (5-7) to understand the molecular epidemiology and reconstruct the phylogeny of M. tuberculosis. The first step was to demonstrate that the chosen molecular methods differentiated isolates with respect to their molecular clonality. The second step was to demonstrate that under proper epidemiologic circumstances, clonality implies active transmission. To eliminate sample bias, we used all data for Guadeloupe during a 3-year period (1994-96) and for Martinique and French Guiana during a 2-year period (1995-96). Table 1. Demographic characteristics of Guadeloupe, Martinique, and French Guiana ------------------------------------------------------------------------- Data Guadeloupe Martinique French Total Guiana ------------------------------------------------------------------------- Surface area (km2) 1,705 1,100 91,000 93,805 Population (1990) 387,034 359,579 114,808 861,421 Population (1996) 422,290 388,340 151,780 962,410 Density (inhabitants 248 353 2 10 /km(sup 2)) Birth rate/1,000 inhabitants 16.8 14.4 29.2 17.8 AIDS(sup a) 700 383 561 1,644 HIV-positive (%)(sup b) 1.4 0.8 1.8 1.3 ------------------------------------------------------------------------- (sup a)Cumulated cases on June 30, 1996. (sup b)On the basis of 7,087 serologic tests performed in departmental clinics in 1994. M. tuberculosis may have developed through the domestical of cattle, and if so, M. bovis was its most probable precursor (8). On the basis of restricted allelic diversity, the speciation of M. tuberculosis is estimated to have occurred 15,000 to 20,000 years ago (4). It is thought that the spread of tuberculosis (TB) began in Europe around the middle ages and reached the new world in the 16th century, Africa in the 19th century, and only recently, remote regions such as Papua-New Guinea (mid-20th century) and the Amazon (last quarter of the 20th century) (8). The study of the genetic biodiversity of M. tuberculosis might help reconstruct this evolutionary scenario. Previous work from our laboratory showed that a significant proportion of patients had conserved (or ancient) strains of tubercle bacilli (6), and analysis based on multiple genetic markers showed genetic relatedness between some clusters of bacilli within Guadeloupe (9). Demographics and Molecular Typing Methods Guadeloupe, Table 2. Epidemiologic and clinical data for 410 Martinique, and tuberculosis (TB) cases reported to the health French Guiana authorities from 1994 to 1996 in the French have major West Indies demographic --------------------------------------------------- differences (Table 1). Guade- Marti- French Total loupe nique Guiana Guadeloupe and --------------------------------------------------- Martinique are densely populated TB cases 136 89 185 410 islands in the Pleuropulmon 115 71 156 342 center of Lesser -nary cases Antilles; they Total TB 11 8.2 40.6 14.20 have similar incidencea population sizes Sex ratio 1.7 1.9 1.5 1.65 (417,000 and PDR to INH (%) 4.4 2.2 3.5 3.4 388,000) and PDR to RIF (%) 1.4 - - - areas (1,705 (km(sup 2) Foreign-born TB 24.5 20.3 68 43.2 and 1,100 km(sup 2) patients (%) and homogeneous TB-HIV coin- 28 19 25 24.6 populations (93% fection (%)b French nationals). --------------------------------------------------- French Guiana is (sup a)New cases/100,000 inhabitants, 1994 to 1996. sparsely (sup b)Total number of known HIV-positive cases among populated TB patients was 101. However, foreign-born patients (150,000), its represented 50% of all the TB-HIV coinfected patients inhabitants in Martinique, 60% in Guadeloupe, and 80% in French scattered Guiana. The % shown is a minimal estimation as HIV throughout a serology results were available only for 49% of large territory patients in Martinique, 75% in Guadeloupe, and 80% (91,000 km2). in French Guiana. Guadeloupe and PDR, primary drug resistance; INH, isoniazid; RIF Martinique are rifampin. characterized by an insular population, whereas French Guiana, which borders Surinam and Brazil, is populated by persons of diverse geographic and ethnic origin (many immigrants from South and Central America). The distribution of TB cases among these three French overseas territories reflects their individual demographics. In Guadeloupe and Martinique, 27% and 25% of all TB cases, respectively, were imported, while in French Guiana, 68% of all cases were imported (10). The basic epidemiologic data (incidence, age, sex, nationality, HIV coinfection) for all TB cases reported to health authorities describe a 3-year period (1994-96) (Table 2). Of 136 TB cases reported in Guadeloupe, 107 had smear- or culture-positive isolates, with culture available in 100 cases; we typed 95 (95%) of 100 isolates. Of 52 cases reported in Martinique, 41 had smear- or culture-positive isolates, and a culture was available in 34 cases. We typed 31 (91%) isolates. Of 124 cases reported in French Guiana, 96 had culture- or smear-positive isolates, with a culture available in 76 cases; we typed 73 (96%) of 76 isolates, along with three that grew in 1995 but came from pathologic specimens received in 1994. The isolates were first studied by spoligotyping (11), which is based on polymorphism of the DR locus of M. tuberculosis (12); spoligotyping results were analyzed by using the Recognizer and Taxotron software (Institut Pasteur, Paris). The 1-Jaccard Index was calculated, and strains were compared by the unweighted pair-group method using arithmetic averages (UPGMA) (13) or by the neighbor-joining method (14). Secondary typing was performed by DRE-PCR (15) or IS6110-RFLP (Table 3) (5). To determine clonality between isolates, we used the following algorithm: strains were considered clonal only when a combination of spoligotyping plus IS6110-RFLP or DRE-PCR indicated they were identical. Data from DR-RFLP (12) and SmaI-RFLP that used a polymorphic GC-rich probe (PGRS) (16) and available epidemiologic information are also included in Table 3. Table 3. Molecular fingerprinting and epidemiologic information from spoligotyping-defined clusters shown in Figure 7.(sup a) ---------------------------------------------------------------------------- No. of strains typed Summary of by results Spoligo- No. of obtained by Clinical and type strains IS6110 DR DRE molecular epidemiologic data, no. harboring RFLP RFLP PCR typing origin, meth. 1 this type meth.2 meth.3 meth.4 methods(sup b) observations ---------------------------------------------------------------------------- 1 2 2 2 2 Different by Found in Surinam and meth. 2, Guadeloupe Beijing IS6110 pattern 2 9 9 2 8 No subcluster 2 patients from by PGRS-RFLP: hospital A and 2 all strains from B + 2 patients identical by with same surnames meth. 2 and 4 IS-type J and by PGRS-RFLP 3 3 3 1 3 2/3 strains Found in Surinam and identical by Guadeloupe, 2 meth. 2, 3, patients from and 4, IS-typehospital A P 5 2 1 1 2 2/2 strains Found in Martinique identical by and Guadeloupe, no meth. 2 and 4,evident IS type not epidemiologic link yet defined 12 2 2 2 1 One spacer Very common pattern, difference represent both with type 14, active transmission identical by and reactivation meth. 2 and 3, IS-type A 13 2 2 2 ND One spacer Suspicion of difference cross-contamination with type 14, (sampling in the identical by same hospital in 3 meth. 2 and 3,days) IS-type A 14 15 15 13 8 Identical by Very common pattern, meth. 2 and 3,represent both identical by active transmission method 4 (one and reactivation band), IS-type A 15 2 2 2 2 2/2 strains Imported cluster identical by (Surinam or meth. 2, 3, Dominican Republic) and 4, IS-type C 17 6 6 1 5 Subclustered 17B found in 2 by PGRS-RFLP: Guadeloupean 17B, 2 strainspatients identical by hospitalized in same meth. 2, 3 andhospital B 4, IS-type N 29 5 5 4 5 5/5 strains 3 of 5 Guadeloupean identical by patients meth. 2, 3, hospitalized in same and 4, IS-typehospital B B 30 2 2 2 2 2/2 strains 2 patients from the identical by same part of meth. 2, 3, Guadeloupe and 4 31 2 2 1 1 2/2 strains 2 Found in French different by Guiana and in meth. Guadeloupe, no epidemiologic link 33 2 ND ND 2 2/2 strains Found in French different by Guiana, no meth. 4 epidemiologic link 34 2 2 1 2 2/2 strains Patients from French identical by Guiana, suspected to meth. 4, be epidemiologically method 2: linked inconclusive, under investigation 42 3 3 1 3 3/3 strains Found in Guadeloupe different by (one patient) and meth. 2 and 4 French Guiana (2 patients), no link 44 2 2 ND 2 2/2 strains 2 patients from St. identical by Maarten (couple) meth. 2 and 4 45 6 1 ND 4 4/6 strains 5 patients from identical by Martinique, one from meth. 4. (2: Guadeloupe, under pending) investigation 46 3 1 ND 3 2/3 strains Found in a identical by Martinique and a meth. 4 Guadeloupe patient, under investigation 50(sup c) 29 16 8 20 Subclustered Imported clusters by meth. 2, 4 (Haïti), other links and PGRS-RFLP,under investigation IS-type E (3 pat.) and F (2 pat) 51 4 3 2 3 2/4 strains Imported cluster identical by (Haïti) for 2 meth. 2 and 4 patients, other (2: pending) links under investigation 53c 29 10 8 20 Subclustered Patients from by meth. 2, 4 cluster T come from and PGRS, the same ward of IS-type K (2 hospital A, 1996 pat.) and T (3 pat.) 61 2 2 1 1 2/2 strains Found in French identical by Guiana in PGRS-RFLP, IS Guadeloupe, under results: underinvestigation investigation 63 2 2 1 2 2/2 strains 2 patients from identical by hospital A in meth. 2 and 4,Guadeloupe, under IS-type R investigation 64 1 1 ND 1 *** 2 isolates from one single patient 65 2 ND ND ND (Results Found in French pending) Guiana, under investigation 66 2 2 ND 2 2/2 strains Found in French identical by Guiana, same surname meth. 2 and 4 67 2 1 ND 2 2/2 strains Found in French identical by Guiana, under meth. 2 and investigation PGRS-RFLP 68 2 1 ND ND (Results Found in one pending) Martinique patient and in Barbados, under investigation ---------------------------------------------------------------------------- (sup a)Of 218 isolates typed, 145 isolates were grouped in 27 distinct spoligo-defined clusters, which were further analyzed by one or more typing methods—IS6110-RFLP (meth. 2), DR-RFLP (meth. 3), and DRE-PCR (meth.4), and sometimes PGRS-RFLP when DRE-PCR or IS6110-RFLP results were inconclusive or unavailable. (sup b)Isolates with matching spoligotypes and matching IS6110 patterns (meth. 1 and 2) or with matching spoligotypes and matching DRE-PCR patterns (meth. 1 and 4) were considered to make up a cluster of epidemiologically associated strains. (sup c)Noninformative spoligotype patterns that lack any discriminating power. *** Two clinical isolates from a single patient. RFLP, restriction fragment length polymorphism; DR, direct repeat; DRE, double-repetitive element; PCR, polymerase chain reaction; PGRS, polymorphic GC-rich probe; ND, not done. Fig. 1. Distribution of tuberculosis cases reported to health authorities in Guadeloupe, Martinique, and French Guiana, [fig] 1994-96, by nationality. DomR = Dominican Republic; Dom = Commonwealth of Dominica. TB Epidemiology Epidemiologic data for the three territories have been described in detail (Table 2) (10). The mean incidence of cases per 100,000 inhabitants for the period studied was 11 in Guadeloupe, 8.2 in Martinique, and 40.6 in French Guiana. The sex ratio varied from 1.5 in French Guiana to 1.9 in Martinique. In each territory, the highest proportion of TB cases was in persons more than 65 years old; most were reactivated infections. The highest number of cases was in adults 25 to 44 years old; most were new cases due to active disease transmission and the high rate of TB-HIV coinfections. Pleuropulmonary disease was most common (80% to 85% of cases), and drug resistance was rare; primary resistance to isoniazid was 2.2% to 4.4%; single drug resistance to rifampin was not observed in Martinique and French Guiana and was limited to 1.4% of cases in Guadeloupe. Four cases of multidrug-resistant TB (MDR-TB, defined as resistance to at least isoniazid and rifampin) were diagnosed; these cases of secondary drug-resistance were caused by noncompliance to standard antituberculosis chemotherapy. In addition, the rate of TB-HIV coinfection in these three territories was high (19% of total TB cases in Martinique, 25% in French Guiana, and 28% in Guadeloupe). Coinfected patients were most likely men 25 to 44 years of age and of foreign origin (1 of 2 TB-HIV coinfected patients was a foreign national in Martinique, compared with 6 of 10 in Guadeloupe and 8 of 10 in French Guiana). These figures agree with the epidemiologic data concerning the distribution of patients on the basis of their nationality (Figure 1). [fig] In the Fig. 2. Nomenclature of the spoligotype database (from 1 to last 20 69) based on published spoligotypes (n = 393) and on years, TB spoligotypes generated during this investigation (n = 218). incidence The corresponding hybridization patterns for oligonucleotides has 1 to 43 (black square, hybridizing; empty square, decreased nonhybridizing) are shown. Type 1 is unique for the Beijing considerably type pattern, whereas type 69 is unique for the Manila type in these pattern. Bold characters illustrate patterns that have so far territories. been noticed only in Caribbean and neighboring Central For American isolates (specific types). Unbolded characters example, illustrate patterns common to those reported elsewhere the (ubiquitous types), whereas italicized characters with an incidence asterisk illustrates patterns not yet found in the Caribbean. of new BCG and H37Rv spoligotypes are shown as controls. cases in Guadeloupe, 36 per 100,000 inhabitants in 1975, declined to 10 per 100,000 inhabitants in the late 1980s. If the same trend continued, TB would vanish by 2000; however, this decline has slowed and incidence has remained at 10 to 12 per 100,000 since 1989. This reversal of decline is linked to poor economic conditions, increase in unemployment, increase in drug and alcohol abuse, and persistent immigration from countries with a high incidence of TB and ongoing HIV epidemics. The Spoligotype Database To define predominant genotypes and trace the origin of strains and their potential movement, we compared spoligotypes of Caribbean isolates with those of other geographic regions. We built a database of 610 spoligotypes (218 from our own investigation and 392 from other countries) with Excel software. The database contains 167 patterns from Goyal et al. (17); 118 from Kamerbeek et al. (11); 106 from Goguet et al. (18); and a single pattern, named the Manila family, from Douglas et al. (19). In this database, 69 spoligotypes shared by more than two patients in any region of the world were numbered types 1 to 69 (Figure 2): types 1-54 denote 395 spoligotypes that were initially ordered from empty to full squares read from left to right, and types 55 to 69 represent 15 new shared types in chronologic order and correspond to 215 spoligotypes investigated later. The nomenclature is temporary until international guidelines are established. Fourteen shared types might be specific for the Caribbean and bordering Central American regions (patterns 5, 12, 13, 14, 15, 17, 29, 30, 54, 63, 64, 66, 67, 68 [bolded in Figure 2]); 23 shared types were found both in the Caribbean region and in other parts of the world (in regular font in Figure 2). The remaining 32 patterns were not present in the Caribbean (highlighted by an asterisk in Figure 2). Fig. 3. A dendrogram represents spoligotyping results of 95 Mycobacterium tuberculosis isolates from Guadeloupe (shared patterns are shown by bold characters). From top to bottom; type 53, ubiquitous; type 44, described by Goguet et al. (two cases) (18); type 54, which does not appear, is found only in isolate 94142, a pattern also found in French Guiana; type 37, which also does not appear, described by Kamerbeek et al. (three cases) (11); type 50, ubiquitous; type 63, specific; type 61, which does not appear is found only in isolate 96131, a pattern also found in Barbados and French Guiana; type 51, [fig] described by Goguet et al. (one case) (18); isolate 96095, a M. bovis BCG clinical isolate; types 12 to 14, specific types; type 5, which does not appear, is found only in isolate 95068, a pattern also found in Martinique; type 31, which does not appear, is found only in isolate 94112, a pattern also found in French Guiana and already reported by Goguet et al. (two cases) (18); type 46, which does not appear, is found only in isolate 95087, a pattern also found in Martinique and was already reported (18); type 17, specific; type 30, specific; types 45 and 15, which do not appear, are found only for isolates 96129 and 94127, respectively, and share patterns with isolates in Martinique and Surinam; type 29, specific; type 2, previously reported by Goguet et al. (one case) (18); type 3, also found in Surinam, has been described by Kamerbeek et al. (one case) (11); type 1, which does not appear, is found only in 95076, a pattern also found in Surinam, and described (20) as the Beijing type. Fig. 4. A dendrogram illustrating spoligotyping results of 76 Mycobacterium tuberculosis isolates from French Guiana (shared patterns are shown in bold). Top to bottom; patterns 50 and 53, ubiquitous; types 54 and 36, which do not appear, are found only for isolates IPC99 and IPC57, respectively; type 34, a ubiquitous type that was limited [fig] to French Guiana in this study; types 66 and 67, specific; type 42, ubiquitous; type 17, specific; type 33, ubiquitous; type 51, ubiquitous; type 31, which does not appear, is found only for isolate IPC23 and is shared with 94112 in Guadeloupe; type 2, shared by four isolates in Guadeloupe. Population Structure of M. tuberculosis in Guadeloupe Of 95 isolates from Guadeloupe (Figure 3), 34 were a unique spoligotype, and 61 shared 13 patterns (six patterns were shared by only two isolates, and seven patterns were shared by the remaining 49 [3 to 13 isolates per pattern]). Patterns 2 (4 isolates), 14 (13 isolates), 29 (5 isolates), 50 (10 isolates), and 53 (12 isolates) were the major shared patterns from Guadeloupe and made up 72% of clustered isolates. The interpretation of the population structure from Guadeloupe shows the presence of important nodes on the dendrogram. As illustrated in Figure 3 (from bottom to top), three distinct patterns show strain 95076 (designated type 1 in our database and identical to the Beijing type commonly found in Asia [20]), the recently reported pattern 3 (11), and pattern 2 (18), respectively. Above the patterns 1-3 (pattern 1 is not shown) lies the previously undescribed pattern 29, which might be specific to our region. Two very different groups can be seen on the next node (the "lower" and "upper" groups). The "lower" group, which comprises 20 isolates and two shared patterns (17 and 30), shows a stepwise polymorphism. This group contains many ubiquitous patterns (reported from at least three geographic regions) shared by a variety of isolates from around the world and concerns isolates 95016, 94105, and 94041 of shared types 20, 42, and 47, respectively. The "upper" group, which comprises the remaining 63 isolates and eight shared spoligotypes (types 12, 13, 14, 51, 63, 50, 44, 53), can be further divided into two subgroups: "upper ubiquitous," which lies above M. bovis BCG strain 96095 and the "upper specific," which lies below the BCG strain (Figure 3). The "upper specific" subgroup is homogeneous (shared types 12, 13, 14, and strain 95068 of shared type 5) and probably has been present in Guadeloupe for a long time (except in a single isolate with pattern 5 from the neighboring island of Martinique). The "upper ubiquitous" subgroup seems considerably closer to M. bovis BCG than do other isolates of M. tuberculosis from Guadeloupe. Population Structure of M. tuberculosis in French Guiana The first population structure of M. [fig] tuberculosis isolates from French Guiana is shown in Figure 4. Of 76 isolates, 30 had a unique spoligotype. Fig. 5. A dendrogram illustrating The remaining 46 shared a total of 11 31 spoligotyping results of patterns: eight patterns (51, 33, 17, Mycobacterium tuberculosis isolates 42, 67, 66, 65, 34) were shared by from Martinique. From top to only two isolates and three major bottom, types 50 and 53, patterns (pattern 2, 4 isolates; ubiquitous; types 52 and 49 are pattern 53, 9 isolates; pattern 50, 17 represented by a single isolate, isolates) were shared by the remaining respectively, M35 and M30, and are 30 isolates. The three major patterns ubiquitous; isolates M10 and M7 represented 65% of clustered isolates belong to specific types 17 and 68, in French Guiana. Most clusters were respectively found in Guadeloupe common to those found in Guadeloupe, and Barbados; M25, belongs to as well as other regions of the world specific type 5 observed in (types 2, 17, 33, 34, 42, 50, 51, 53, Guadeloupe; type 45, ubiquitous; 65), except two patterns that have type 46, ubiquitous; isolate M23 been so far only reported from French shares type 2, with isolates in Guiana (types 66, 67). No isolate was Guadeloupe and French Guiana. of the Beijing type, which is unexpected, considering the number of persons of Chinese origin in French Guiana. A high degree of heterogeneity was observed in French Guiana, which is not surprising given the large surface area, high number of immigrants and persons of various ethnic origins, and high TB incidence rate. Population Structure of M. tuberculosis in Martinique Of 31 isolates studied in Martinique (Figure 5), 19 had an unshared spoligotype, and 12 shared 4 patterns (pattern 46, 2 isolates; pattern 45, 5 isolates; pattern 53, 2 isolates; pattern 50, 3 isolates). All the clusters in Martinique were also found in Guadeloupe (45, 46, 50, and 53). Pattern 45 is overrepresented in this population, which could suggest active transmission of this clone of tubercle bacilli in Martinique. Other patterns common in Guadeloupe are only poorly represented in Martinique: type 2 (isolate M23), type 5 (isolate M25), and type 17 (isolate M10). Type 5 is specific only to Martinique and Guadeloupe. On the contrary, Martinique shares some patterns with the rest of the world that are not found in Guadeloupe or French Guiana (type 49, isolate M30; type 52, isolate M35). Despite the small sample size of isolates from Martinique, many isolates (clustered or unclustered) share published patterns from the rest of the world (18 of 31 isolates studied). Population Structure of M. tuberculosis in Other Caribbean Regions We also investigated 16 additional M. [fig] tuberculosis isolates (12 from Surinam and 4 from Barbados). Although not Fig. 6. A dendrogram representative of the population studied, illustrating spoligotyping spoligotyping of these isolates allowed results of 16 Mycobacterium detection of some interesting links tuberculosis clinical (Figure 6). Fourteen unique types and one isolates from Barbados and shared type (pattern 53, two isolates) Surinam. From top: type 53, were found. Among the unclustered ubiquitous; isolates Barb.3 isolates, strain 94030 from Surinam and 94030 belong to specific matched isolate 94127 from Guadeloupe types 68 and 15, (type 15), strain 94018 from Surinam respectively; isolates 94018, matched type 19 from Holland (11), and 94020, 94034, and Barb.1 strain 94034 from Surinam matched isolate belong to ubiquitous types 95047 (type 3) from Guadeloupe. 19, 1, 3, and 61, Furthermore, two strains from Barbados respectively. shared patterns with isolates from other geographic areas: Barb 3 to M7 (type 68) from Martinique and strain Barb 1 to the pattern 61 from France (18). Predominant Genotypes and Strain Origin and Transmission The dendrogram representing the global structure of the population studied (Figure 7) shows the potential historical or epidemiologic interregional flux of M. tuberculosis between Caribbean and neighboring Central American regions. Of the 218 isolates, 145 (66.5%) had 29 shared types. Twenty-five isolates were not initially grouped but were clustered only when a dendrogram of all 218 isolates was drawn. These interregional links defined eight new types in the database (patterns 1, 5, 15, 31, 54, 61, 64, and 68) for 15 isolates; the remaining 10 isolates enriched patterns already present (Figure 7). These links, made by finding clonal strains in distant geographic territories, do not necessarily define epidemiologic relationships between these strains (21) but are unlikely to be due to independent genetic evolution. The dendrogram is an indirect picture of the common history of TB spread in this part of the world. For example, spoligotype 2 was recently proposed to have originated in Latin America (22). Our recent investigations favor this hypothesis as we have traced this pattern in all three territories investigated (Martinique [one isolate]), Guadeloupe [four isolates], and French Guiana [four isolates]). Isolates from all nine patients were further investigated by IS6110-RFLP, DRE-PCR, and SmaI-PGRS typing (23), and the results confirmed the clonality of this specific cluster. Searching for this genotype in other Latin American countries may help elucidate its origin and distribution. To define the clonality of the isolates and to be in accordance with the current practice for [fig] molecular typing of M. tuberculosis using spoligotyping as a first-line method (7, 11, 17, 18), we systematically performed a study of a Fig. 7. A cumulative spoligotyping-defined cluster by a second method dendrogram for the 218 (IS6110-RFLP or DRE-PCR). When corroborated by Caribbean isolates of epidemiologic information, the clusters so Mycobacterium defined were considered evidence of ongoing tuberculosis. New types active transmission of TB (Table 3). However, in not visible on individual numerous cases, the epidemiologic data were not dendrograms can now be conclusive, and molecular clonality was not observed (types 1, 5, 15, considered direct proof of a link. Thus, despite 31, 54, 61, 64, and 68). suspected links in 81 (37%) of 218 of cases on the basis of molecular typing alone, a direct epidemiologic link was demonstrated in 31 (14.2%) of 218 of cases. However, future prospective studies including active case-finding and source tracing may help increase the number of epidemiologically linked cases in our region. Distribution of Spoligotypes Bearing Geographic Specificities To reconstruct the evolutionary tree of M. tuberculosis on the basis of 610 different spoligotypes (mostly from the United Kingdom, Holland, France, the Caribbean, and the neighboring Central American region), we performed a similarity search (Figure 2). Although the tree may not reflect the full diversity of spoligotypes, it suggests the existence of distinct families of spoligotypes with geographic specificities (Figure 8). Some strain families may be related to specific populations, geographic regions, and the history of TB spread. For example, the Beijing genotype, which is most divergent in the tree shown in Figure 8 (type 1), would have undergone the most extensive genetic evolution since the origin of M. tuberculosis. This information is consistent with the mechanism of evolution of the DR locus, which appears to proceed through losses of direct repeats (24,25). Although in an unrooted tree, the positions of patterns 50 and 53 correlate well with the isolates most widely represented internationally (116 of 610 isolates studied; 54 of 218 isolates from the Caribbean and neighboring Central American region and 62 of 392 published spoligotypes from different parts of the world) and may constitute a candidate root for the interpretation of distances between isolates. Genetic Evolution of Tubercle Bacilli The genetic evolution of tubercle [fig] bacilli is closely associated with the past and present of its host. Fig. 8. A preliminary Consequently, human migration, phylogenetic tree obtained by population mixing, and other the neighbor-joining algorithm sociodemographic factors have long on the basis of the 1-Jaccard played an important role in the spread index (Sj=a/a+c, where a is the and subsequent evolution of the M. number of simultaneously tuberculosis genome. The insular model positive characters and c is the of the Caribbean, in which human number of discrepancies), which migration (hence the introduction of the is not exhaustive for all disease) was estimated to occur only existing phylogenetic links. A approximately 400 years ago, is total of 70 shared spoligotypes particularly interesting for discovering were analyzed (69 types shown in conserved strains of tubercle bacilli Figure 2 and Mycobacterium bovis and for detecting rare M. tuberculosis BCG). genotypes. In this context, the DR locus is a unique chromosomal region characteristic of the M. tuberculosis complex, which shows a high degree of polymorphism that involves homologous recombination and IS-mediated transposition (24). The high degree of internal homology within the DR region of M. tuberculosis is likely to favor such genetic rearrangements. Similarly, transposition of IS6110 is instrumental in generating new subclones of M. tuberculosis (26). Following the principle of Dollo parsimony, which assumes that losses of genes are much more common and likely than independent evolutionary origins (27), the evolutionary process of M. tuberculosis can be speculated to have involved the loss of DR repeats. However, the tree in Figure 8 does not perfectly represent all phylogenetic links between isolates as the mechanisms of loss of DR elements (by homologous recombination or replication slippage) could involve simultaneous loss of >1 of the 43 building blocks. The present phylogenetic analysis will be extended to other multiple genetic markers to include variables recently proposed by Sreevatsan et al. (4). However, construction of such trees on the basis of the simultaneous feeding and computer analysis of multiple mycobacterial markers remains cumbersome and constitutes a priority research topic for the studies of M. tuberculosis genome evolution. --------------------------------------------------------------------------- Acknowledgments The authors thank F. Pfaff, L. Schlegel, P. Levett, and P. Prabhakar for sending the cultures for identification to the Institut Pasteur, Guadeloupe, and F. Prudenté, M. Berchel, and K.S. Goh for their expert technical help. This project was financed by research grants provided by Délégation Générale au Réseau International des Instituts Pasteur et Instituts Associés, Institut Pasteur, Paris, and Fondation Française Raoul Follereau, Paris, France. Dr. Sola is Chargé de Recherche at the Tuberculosis and Mycobacteria Unit at the Pasteur Institute of Guadeloupe, which is headed by Dr. Rastogi. The authors work closely with public health agencies in Guadeloupe, Martinique, and French Guiana to support the tuberculosis control program. They provide expertise and training on the genetic characterization and laboratory diagnosis of mycobacteria. Their current research interests include molecular diagnostics and epidemiology, drug resistance, and mechanisms of mycobacterial pathogenicity. Address for correspondence: Nalin Rastogi, Unité de la Tuberculose & des Mycobactéries, Institut Pasteur, Morne Jolivière, B.P. 484, F-97165 Pointe à Pitre-Cedex, Guadeloupe; fax: 590-893-880; e-mail: rastogi@ipagua.gp. References 1. Cole ST, Brosch R, Parkhill J, Garnier T, Churcher C, Harris D, et al. Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 1998;393;537-44. 2. van Helden PD. Bacterial genetics and strain variation. Novartis Foundation Symposium 1998;217:178-94. 3. Small PM, van Embden JDA. Molecular epidemiology of tuberculosis. In: Bloom BR, editor. Tuberculosis: pathogenesis, protection and control. Washington: American Society for Microbiology; 1994, p. 569-82. 4. Sreevatsan S, Pan X, Stockbauer KE, Connell ND, Kreiswirth BN, Whittam TS, et al. Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent global dissemination. Proc Natl Acad Sci U S A 1997;97:9869-74. 5. van Embden JDA, Cave MD, Crawford JT, Dale JW, Eisenach KD, Gicquel B, et al. Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: recommendations for a standardized methodology. J Clin Microbiol 1993;31:406-9. 6. Sola C, Horgen L, Goh KS, Rastogi N. Molecular fingerprinting of Mycobacterium tuberculosis on a Caribbean island with IS6110 and DRr probes. J Clin Microbiol 1997;35:843-6. 7. Sola C, Horgen L, Maïsetti J, Devallois A, Goh KS, Rastogi N. Spoligotyping followed by double-repetitive element PCR as rapid alternative to IS6110-fingerprinting for epidemiological studies of tuberculosis. J Clin Microbiol 1998;36:1122-4. 8. Stead WM, Eisenach KD, Cave MD, Beggs ML, Templeton GL, Thoen CO, et al. When did Mycobacterium tuberculosis infection first occur in the New World? Am J Respir Crit Care Med 1995;151:1267-8. 9. Sola C, Horgen L, Devallois A, Rastogi N. Combined numerical analysis based on the molecular description of Mycobacterium tuberculosis by four-repetitive sequence-based DNA typing sequence. Res Microbiol 1998;149:349-60. 10. Rastogi N, Schlegel L, Pfaff F, Jeanne I, Magnien C, Lajoinie G, et al. La tuberculose dans la Région Antilles-Guyane: situation epidemiologique de 1994 à 1996. Bulletin Epidémiologique Hebdomadaire 1998;11/98:45-7. 11. Kamerbeek J, Schouls L, van Agterveld M, van Soolingen D, Kolk A, Kuijper S, et al. Simultaneous detection and strain differentiation of Mycobacterium tuberculosis for diagnosis and epidemiology. J Clin Microbiol 1997;35:907-14. 12. Hermans PWM, van Soolingen D, Bik EM, de Haas PEW, Dale JW, van Embden JDA. Insertion element IS987 from Mycobacterium bovis BCG is located in a hot-spot integration region for insertion elements in Mycobacterium tuberculosis complex strains. Infect Immun 1991;59:2695-705. 13. Sneath PHA, Sokal RR. Numerical taxonomy: the principles and practices of classification. San Francisco (CA): W.H. Freeman & Co.; 1973. 14. Saitou N, Nei M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 1987;4:406-25. 15. Friedman CR, Stoeckle MY, Johnson WD, Riley LW. Double-repetitive-element PCR method for subtyping Mycobacterium tuberculosis clinical isolates. J Clin Microbiol 1995;33:1383-4. 16. Poulet S, Cole ST. Characterization of the highly abundant polymorphic GC-rich repetitive sequence (PGRS) present in Mycobacterium tuberculosis. Arch Microbiol 1995;163:87-95. 17. Goyal M, Saunders NA, van Embden JDA, Young DB, Shaw RJ. Differentiation of Mycobacterium tuberculosis isolates by spoligotyping and IS6110 restriction fragment length polymorphism. J Clin Microbiol 1997;35:647-51. 18. Goguet de la Salmonière Y, Li HM, Torrea G, Bunschoten A, van Embden JDA, Gicquel B. Evaluation of spoligotyping in a study of the transmission of Mycobacterium tuberculosis. J Clin Microbiol 1997;35:2210-4. 19. Douglas JT, Qian L, Montoya JC, Sreevatsan S, Musser JT, van Soolingen D, et al. Detection of a novel family of tuberculosis isolates in the Philippines. 97th general meeting of the American Society for Microbiology, Washington: American Society for Microbiology Press; 1997. p. 572. 20. van Soolingen D, Qian L, de Haas PEW, Douglas JT, Traore H, Portaels F, et al. Predominance of a single genotype of Mycobacterium tuberculosis in countries of East Asia. J Clin Microbiol 1995;33:3234-8. 21. Yang Z, Barnes PF, Chaves F, Eidenach KD, Weis SE, Bates JH, et al. Diversity of DNA fingerprints of Mycobacterium tuberculosis in the United States. J Clin Microbiol 1998;36:1003-7. 22. Prodinger WM, Pavlic M, Pedroza JC, Allenberger FJ. Tracing of a cluster of tuberculosis infections [abstract U-56]. 98th General Meeting of the American Society for Microbiology, Washington: American Society for Microbiology Press; 1998. p. 504. 23. Jasmer RM, Ponce de Leon A, Hopewell PC, Alarcon RG, Moss AR, Paz A, et al. Tuberculosis in Mexican-born persons in San Francisco: reactivation, acquired infection and transmission. Int J Tuberc Lung Dis 1997;1:536-41. 24. Groenen PMA, Bunschoten AE, van Soolingen D, van Embden JDA. Nature of DNA polymorphism in the direct repeat cluster of Mycobacterium tuberculosis; application for strain differentiation by a novel typing method. Mol Microbiol 1993;10:1057-65. 25. van Soolingen D, de Haas PEW, Hermans PWM, Groenen PMA, van Embden JDA. Comparison of various repetitive DNA elements as genetic markers for strain differentiation and epidemiology of Mycobacterium tuberculosis. J Clin Microbiol 1993;31:1987-95. 26. Bifani PJ, Plykaitis BB, Kapur V, Stockbauer K, Pan X, Lutfey ML, et al. Origin and interstate spread of a New York City multidrug-resistant Mycobacterium clone family. JAMA 1996;275:452-7. 27. Meyer A. In: Harvey PH, Leigh Brown AJ, Maynard Smith J, editors. New uses for new phylogenies. London: Oxford University Press; 1996. p. 322-40. ------------------------------------------------------------------------ Research Human Rabies Postexposure Prophylaxis during a Raccoon Rabies Epizootic in New York, 1993 and 1994 Jeffrey D. Wyatt,* William H. Barker,* Nancy M. Bennett,† and Cathleen A. Hanlon‡ *University of Rochester School of Medicine & Dentistry, Rochester, New York, USA; †Monroe County Department of Health, Rochester, New York, USA; ‡Centers for Disease Control & Prevention, Atlanta, Georgia, USA --------------------------------------------------------------------------- We describe the epidemiology of human rabies postexposure prophylaxis (PEP) in four upstate New York counties during the 1st and 2nd year of a raccoon rabies epizootic. We obtained data from records of 1,173 persons whose rabies PEP was reported to local health departments in 1993 and 1994. Mean annual PEP incidence rates were highest in rural counties, in summer, and in patients 10 to 14 and 35 to 44 years of age. PEP given after bites was primarily associated with unvaccinated dogs and cats, but most (70%) was not attributable to bites. Although pet vaccination and stray animal control, which target direct exposure, remain the cornerstones of human rabies prevention, the risk for rabies by the nonbite route (e.g., raccoon saliva on pet dogs' and cats' fur) should also be considered. Raccoon rabies, present in the southeastern United States since the 1950s, became responsible for an epizootic in the U.S. mid-Atlantic region during the 1970s after raccoons were translocated there for hunting (1). The introduction of the variant of rabies virus associated with raccoons into a rabies-naive raccoon population caused the most intensive animal rabies outbreak on record, in part because of the abundance of raccoons in suburban environments throughout the mid-Atlantic and northeastern metropolitan areas. Raccoon rabies affects approximately one million square kilometers of the eastern United States with a human population of approximately 90 million. Since the mid-Atlantic raccoon rabies epizootic entered New York State in 1990, the number of rabid animals increased from 54 in pre-epizootic 1989 to 2,746 (89% raccoons) in 1993—the largest number of rabid animals ever reported from any state (2). Despite traditional public health measures for rabies control (e.g., pet vaccination, stray animal control, public education), human rabies postexposure prophylaxis (PEP) rates inevitably increased with the arrival of the epizootic front (3). Preliminary data from New York documented a 4,000% increase in the absolute number of persons receiving PEP, from 81 (1989) to 3,336 (1993) (4). The epidemiologic trends of human PEP in New York State remain largely undescribed. One of the Healthy People 2000 objectives formulated by the U.S. Public Health Service is to reduce by 50% the need for human rabies PEP by the year 2000 (5). A reduction in the number of PEP cases, which are not reportable, appears unattainable without first defining the numerator, as well as the epidemiologic characteristics of precipitating events leading to suspected rabies exposure and inappropriate treatments. We describe demographic and animal exposure data associated with human rabies PEP in an area with epizootic raccoon rabies. The epidemiologic description is intended to assist medical practitioners and public health officials in reducing the incidence of human and domestic animal exposure to rabid animals and thus in minimizing the need for PEP in communities affected by the raccoon rabies epizootic. The Study Setting Four contiguous upstate New York counties (Monroe, Wayne, Cayuga, and Onondaga) were first affected by the raccoon rabies epizootic between December 1992 and June 1993 (Figure 1). Monroe and Onondaga Counties, encompassing the cities of Rochester and Syracuse, are predominantly urban-suburban, with human population densities of 414 per square kilometer and 230 per square kilometer, respectively. Wayne and Cayuga Counties are predominantly rural-suburban, with relatively lower population densities of 57 and 45 people per square kilometer, respectively. The four-county region in western upstate New York comprises 7,090 square kilometers and has an estimated human population of 1,354,377. Data Characteristics and Sources We considered all human rabies PEP cases reported in 1993 and 1994 for study area. The PEP capture rate was believed high because local health units were responsible for providing funds for any treatment expenses not covered by health insurance, and a completed, rabies report form was required before reimbursement of the [fig] local health unit from state funds. The New York State Sanitary Code requires physicians to report potential Figure 1. New York State raccoon rabies human exposure to rabies and PEP epizootic progression (1990-94). administration to county health The raccoon rabies epizootic first departments. We abstracted data from affected Monroe, Wayne, Cayuga, and these standardized reports and patient Onondaga Counties between December records. Data were grouped by patient 1992 and June 1993. demographics, animal characteristics, and exposure details. Exposure source was defined as the suspected- or confirmed-rabid animal that directly or indirectly resulted in one or more potential human exposures to rabies. Direct contact exposure consisted of direct contact (e.g., bite, scratch) or contamination of mucous membranes with potentially infectious material from a rabid animal. Indirect contact consisted of contamination from a fomite (e.g., through racoon saliva on a pet's fur with a pet owner's open wounds or mucous membranes). Analyses Population figures from the 1990 New York State census were used to calculate the incidence of PEP by county, age, and gender (6). Descriptive analyses of data elements were made through queries of Microsoft Access relational database. Each PEP contributed to the denominator of the analyses. Since multiple PEP cases occurred from exposure to a single animal, data for individual animals were also summarized. Chi-square tests were performed with Epi-Info Version 5 software. Findings PEP Incidence The annual PEP incidence for the study area increased from <1 case per 100,000 residents in pre-epizootic 1992 to 35 cases in 1993 and 52 cases in 1994. Of 1,173 cases of human rabies PEP in the study areas, 474 were reported in 1993 and 699 cases were reported in 1994. The mean annual incidence of PEP was 32 cases per 100,000 for the urban counties (Monroe and Onondaga; 315 residents per square kilometer) and 123 cases per 100,000 for the two rural counties (Wayne and Cayuga; 51 residents per square kilometer). Season The number of PEP cases peaked in summer to early autumn (Figure 2). During 1993, the highest number of PEP cases occurred approximately 4 to 6 months (August through November) after the invasion of raccoon rabies during March through June 1993; in 1994, the highest number occurred in summer (June through August). Gender and Age Gender and age data were available for 100% and 95% of all patients, respectively. Of 1,173 PEP cases, 642 (55%) were administered to male and 531 (45%) to female patients. The mean annual incidence of PEP in male and female patients was 47 and 38 per 100,000, respectively. The PEP rates were highest in persons 10 to 14 years of age (165 per 100,000) and 35 to 44 years of age (113 per 100,000) (Figure 3). The median age was 29 and 31 years for male and female patients, respectively. No significant relationship was observed between gender and age groups for the study area. [fig] [fig] Figure 2. Human rabies Figure 3. Human rabies postexposure postexposure prophylaxis in four New prophylaxis in four New York State York State counties (Cayuga, Monroe, counties (Cayuga, Monroe, Onondaga, and Wayne), Onondaga, and Wayne), 1993-1994, 1993-94: by month. incidence by gender and age. Exposure Source Species Exposure to wild animals accounted for 783 (67%) of all PEP cases (Table 1). Among wildlife, raccoons were by far the leading source of exposure, accounting for 589 (75%) of 779 PEP cases due to wildlife exposure. The other sources of wildlife exposure were bats (54 cases), skunks (35 cases), foxes (28 cases), white-tailed deer (13 cases), woodchucks (12 cases), small rodents (9 cases), sika deer (4 cases), and other wild species (39 cases). Of 390 domestic animal exposures resulting in PEP, 205 were attributed to cats, 165 to dogs, 12 to cattle, 5 to pet rabbits, and 3 to horses. Among PEP cases resulting from exposure to cats and dogs, 66% and 67%, respectively, were initiated after contact with stray animals unavailable for the recommended 10-day confinement and observation to rule out rabies or euthanasia and testing. Dog exposures were disproportionately higher in urban (137 [18%] of 753) than in rural counties (28 [7%] of 420) (p <0.001) (Table 2). In urban areas, dog exposure was primarily due to stray or unowned dogs (95 [69%] of 137). In rural areas, stray or unowned dogs accounted for 11 (39%) of 28 dog exposures (p <0.01). Only one dog (in a rural county) tested positive for rabies in the study area in 1993 and 1994. Of 68 pet cats resulting in human exposure, 61 (90%) were not vaccinated against rabies compared with 14 (24%) of 58 pet dogs (p <0.001). Table 1. Human rabies postexposure prophylaxis (PEP), New York State, 1993-94(sup a) ---------------------------------------------------------------------------------- Nonbite (N=818) ------------------------------------------- Direct(sup b) ----------------------------- Bite Indirect(sup c) Overall Animal (N=355) Scratch Saliva NT Blood Saliva total PEP source N (%) N ( %) N (%) N (%) N (%) N (%) N (%) ----------------------------------------------------------------------------------- Raccoon 37 (10) 18 (33) 44(29) 4 (67) 13 (93) 472(sup d)(79) 589(sup e)(50) Bat (all 29 (8) 3 (6) 12 (8) 1 (17) 0 (0) 9 (2) 54 (5) species) Other wild species 24 (7) 5 (11) 21(14) 0 (0) 1 (7) 89 (15) 140 (12) All wild species 90 (25) 26 (47) 77(51) 5 (84) 14(100) 570 (96) 783 (67) Cat 114 (32) 29 (53) 41(28) 0 (0) 0 (0) 21 (4) 205 (17) Dog 151 (43) 0 (0) 13 (9) 0 (0) 0 (0) 1 (<1) 165 (14) Other domestic species 0 (0) 0 (0) 19(13) 1 (17) 0 (0) 0 (0) 20 (2) All domestic species 265(75) 29(53) 73(49) 1 (17) 0 (0) 22 (4) 390 (33) Total 355 (30) 55(55) 150(13) 6(0.5) 14 (1) 592 (51) 1,173 (100) ----------------------------------------------------------------------------------- (sup a)Data are from Cayuga, Monroe, Onondaga, and Wayne Counties. (sup b)Direct contamination of an open wound or mucous membrane with potentially infectious material such as saliva, nervous tissue (NT), or blood (mixed with other body fluids), from a rabies-suspect or known-rabid animal. (sup c)No direct contact with a rabid or suspect-rabid animal. Indirect exposure through possible conveyance of saliva on an animal (i.e., pet dog or cat) or inanimate object resulting in contamination of an open wound or mucous membrane. (sup d)p < 0.001. More people received PEP after indirect exposure to saliva from raccoons than from any other species (472 PEP cases due to indirect contact with 261 raccoons). (sup e)Total PEP cases with raccoon as an exposure source (includes one case with no reported route of exposure). Table 2. Human rabies postexposure prophylaxis Type of (PEP) in New York State, 1993-94: urban and rural Exposure settings(sup a) Of 1,173 ----------------------------------------------------------------- PEP All four cases, Urban Rural counties 355 --------------------------------------- (30%) Animal source N (%) N (%) N (%) resulted from -------------------------------------------------------------- animal Dog(sup b) 137 (18) 28 (7) 165 (14) bites Cat 130 (17) 75 (18) 205 (17) and 817 (70%) Other domestic from species(sup c) 5 (<1) 15 (4) 20 (2) nonbite All domestic encounters species 272 (35) 118 (28) 390 (33) (Table Raccoon 41 (45) 248 (59) 589 (50) 1). A route of Bat (all species) 41 (5) 13 (1) 54 (5) exposure Striped skunk 29 (4) 6 (<1) 35 (3) not Fox 19 (3) 9 (2) 28 (2) reported Other wild in one species(sup d) 51 (4) 26 (2) 77 (7) case involved All wild species 481 (65) 302 (72) 783 (67) a Total 753 420 1,173 raccoon. Rate per Suspected 100,000 pop. 32 123 43 contact with -------------------------------------------------------------- animal (sup a)haracteristics of human rabies PEP cases reported to the saliva health departments of the two relatively urbanized counties, (148 Onondaga and Monroe, and the two relatively rural counties cases Cayuga and Wayne, during 1993 and 1994. from (sup b)p < 0.00. Human PEP rates due to dog exposures were direct significantly higher in urban counties. contact (sup c)Other domestic species include 2 and 3 PEP cases due and 594 to cow from and horse exposure in the urban counties and 10 and 5 cases due indirect to cow and domestic rabbit exposure in the rural counties, contact) respectively. was (sup d)Other wild species includes 17, 6, 4, 2, 2, and 1 PEP responsible cases due to an unknown animal type, wild rodent (other than for 742 woodchuck), 4 Sika deer (exotic, captive species), opossum, (91%) of coyote, and mink in the urban counties and 17 and 3 PEP cases due to due to an unknown animal type and wild rodent (other than nonbite woodchuck) in the rural counties, respectively. 817 PEP cases exposure; contact with nervous tissue (6 PEP cases) or blood (14 PEP cases) accounted for 2% of cases due to nonbite exposure. Fifty-five (7%) of 817 nonbite exposures were attributable to scratches from 23 cats (responsible for 29 PEP cases), 16 raccoons (18 cases), 3 bats (3 cases), 2 wild rodents (2 cases), and 3 other wild animals (3 cases). Of 355 bite exposures, 265 (75%) involved domestic animals (151 due to 150 dogs, 114 due to 108 cats); 90 (25%) involved bites from wild animals—34 raccoons (responsible for 37 PEP cases), 27 bats (29 cases), 9 rodents (9 cases), and 13 other wild species (15 cases). Mode of Contact Of 1,173 cases, 594 (51%) occurred because of possible indirect contact with a suspect rabid animal; 583 (98%) of 594 occurred after suspected exposure to saliva from a rabid (or suspect rabid) animal on the fur of a nonsuspect dog, cat, or other animal. In 9 (2%) of these cases, PEP was administered after suspected exposure by possibly contaminated fomites including door knobs, traps, arrows, a flashlight, and a wire. Possible indirect exposure to dogs with potentially infectious material on their fur resulted in 507 (85%) PEP cases, while suspected indirect exposure by cats resulted in 70 (12%) cases. Other suspected exposure sources were a horse, rabbit, pet duck, chicken, wild bird, captive exotic sika deer, and a person. Group Exposure Exposure of one person to a suspect rabid animal precipitated 625 (53%) PEP cases; the remaining 548 (47%) occurred after more than one person was exposed to the same suspected animal (Table 3). Exposure of a single person was more likely associated with a bite (p <0.001), while group exposure (involving two or more persons) was more likely associated with nonbites (p <0.001). Wild animal species accounted for most group exposures—with three exceptions. The largest group exposures (involving 12, 13, and 14 people) were associated with the handling of rabid domestic animals (before diagnosis) by veterinary clinic employees. Rabies Status The laboratory diagnosis of rabies in the exposing animal was associated with 540 (46%) of all PEP cases (445 due to wildlife and 95 due to domestic animals). Eighty-nine percent of PEP cases attributed to rabid wildlife involved raccoons. In 88 cases, PEP was initiated after contact with animals eventually proven nonrabid. In 544 cases PEP was administered after contact with animals not tested for rabies. Confirmation of rabies in suspect domestic animals occurred in association with 91 (23%) of 390 PEP cases resulting from exposure to domestic species, including 40 due to 5 pet cats, 23 to 5 stray cats, 13 to 1 pet dog, 5 to a domestic rabbit, 7 to a cow, and 3 to a horse. Conversely, in 88 (8%) of all cases PEP was given after encounters with 81 animals subsequently proven nonrabid (35 due to 33 cats, 32 to 32 dogs, 9 to 8 raccoons, 3 to 3 bats, 5 to 2 skunks, 1 to 1 woodchuck, 1 to 1 squirrel, and 2 to 1 muskrat). Table 3. Human rabies postexposure prophylaxis (PEP) in New York State, 1993-94: Epidemiologic characteristics(sup a) --------------------------------------------------------------------------- Group size ------------------------------------------------------ 1 2 3 4 5 6+ ------------------------------------------------------ Characteristic N % N % N % N % N % N % --------------------------------------------------------------------------- Number 625 53 180 15 84 7 112 10 60 5 112 10 No. of sources 625 79 90 11 28 4 28 4 12 2 13 2 Route of exposure Bite(sup b) 328 52 17 9 4 5 3 3 0 0 3 3 Nonbite 296 47 163 91 80 95 109 97 60 100 109 97 Unknown 1 0.2 0 0 0 0 0 0 0 0 0 0 Source of exposure Dog or cat 273 44 16 9 3 4 4 4 10 17 64 57 Other domestic species 2 0.3 4 2 3 4 4 4 5 8 6 5 Raccoon 235 38 124 69 57 68 96 86 35 58 42 38 Bat 39 6 12 7 3 4 0 0 0 0 0 0 Other wild species 76 12 24 13 18 21 8 7 10 17 0 0 Mean age (yr) 33.4 32.5 24.8 22.5 21.8 26.0 --------------------------------------------------------------------------- (sup a)PEP data are from Cayuga, Monroe, Onondaga, and Wayne Counties. (sup b)Probability of bite exposure for PEP involving single person vs. group of >1 PEP cases, p <0.001. Of 540 cases of PEP associated with animals proven to be rabid, 505 (94%) were due to suspected saliva exposure; 22 (4%) and 13 (2%) involved bites or scratches, respectively. Conversely, 71 (81%) of 88 PEP cases associated with nonrabid animals (i.e., laboratory-confirmed as negative or confined and observed to be healthy) occurred after bite exposures. Of the 544 PEP cases associated with animals of unknown rabies status, 48% were due to bites, 45% to suspected saliva contacts, and 7% to scratches. Wild animals accounted for 98% of the 690 animals submitted and testing positive for rabies in the study area for 1993-94; 613 were raccoons. If animals testing positive for rabies are used as a surrogate for the true incidence, an approximately 20-fold increase in PEP cases per rabid domestic animal compared with each rabid wild animal, regardless of rural or urban region, is seen (data not shown). Provoked Exposures A provoked exposure was characterized by intentional, human-initiated interaction with a suspect rabid animal. Cases resulting from provoked exposure accounted for 392 (33%) of 1,173 of all PEP cases; 248 (63%) involved domestic animals. Most cases resulting from provoked exposure of domestic animals involved cats (162 [65%] of 248) and less frequently, dogs (62 [25%] of 248). Wild animals accounted for 144 (37%) PEP cases from provoked exposure. Time of PEP Initiation The interval between exposure to suspect rabid animals and initiation of PEP was 0 to 43 days (median 2 days). Bite exposures were associated with no delay in treatment; nonbite exposures were associated with a 3- to 4-day interval (p <0.001). PEP Regimen In 1993 and 1994, postexposure biologic products licensed for use in the United States were rabies vaccine adsorbed, human diploid cell vaccine (Imovax), and human rabies immune globulin (HRIG; Hyperab or Imogam). As recommended by the Advisory Committee on Immunization Practices (ACIP), PEP for the rabies-naive person consists of HRIG (20 IU/kg) on day 0 and five doses of rabies vaccine administered on days 0, 3, 7, 14, and 28 (7). Scheduling information was unavailable for our cases. Administration of PEP biologic products was recorded as complete in 1,016 (87%) of 1,173 PEP cases. Information regarding completion of the treatment series was not available in 15 cases (1%). Appropriate PEP for preimmunized persons consists of two vaccine doses on days 0 and 3 (7) and was administered to 26 persons, accounting for 2% of all cases. Among preimmunized persons, 17 (65%) of 26 PEP cases occurred after occupational exposures by 11 veterinary staff personnel (including two group exposures to proven rabid cats), four wildlife rehabilitators, one health department employee, and one police officer. In 54 (5%) instances, PEP was discontinued because of lack of clinical signs in 29 dogs (29 PEP cases) and 23 cats (25 PEP cases) confined for the recommended 10-day observation period. Moreover, 34 (3%) PEP cases were discontinued because of rabies-negative laboratory results in 10 cats (10 PEP cases), 7 raccoons (8 cases), 2 skunks (5 cases), 4 dogs (4 cases), 3 bats (3 cases), 1 muskrat (2 cases), 1 woodchuck (1 case), and 1 squirrel (1 case). After PEP was initiated, 29 (2%) of 1,173 refused to complete the series; two cited adverse reactions. In nine cases PEP deviated from ACIP recommendations: apparently inadvertent scheduling and administration of six total vaccine doses in four patients and intentional omission of HRIG in the treatment regimen of five patients. Adverse Effects The categories available for characterizing adverse effects on the state rabies report form were none, slight, moderate, severe, or unknown. In 596 (51%) of 1,173 PEP cases, no information was recorded. Of 577 responses, 495 (86%) reported no adverse effects resulting from PEP. Adverse effects were characterized as slight by 67 (12%) persons. Moderate adverse reactions including vomiting, nausea, fever, aches, and weakness were reported by 13 (2%) persons. Serious systemic adverse reactions, recorded as anaphylactic shock and serum sickness, occurred in two (0.2%) persons. Both of these patients had received HRIG; PEP was discontinued after one and two vaccine doses in each case. Conclusions The most important finding of this study was that in most cases PEP was administered because of suspected nonbite, indirect exposure to animal saliva, a route conventionally thought of nearly negligible risk in rabies transmission (7,8). Because of effective PEP, public health personnel and health-care workers are primarily challenged with the assessment of exposure to rabies, rather than with treatment of human cases of the disease. Assessment of nonbite saliva exposures are particularly time-consuming and should consist of a thorough, but nonleading, history-taking that elicits the probability or confirmation of mucous membrane or nonintact skin contact and a realistic assessment of the potential presence of infectious saliva on surfaces or pets. Given the invariably fatal outcome of clinical rabies, the tendency may be to administer PEP, even without clear indication of exposure. This tendency may be unwise—not only for economic reasons, but also because, despite their relative innocuity and high potency, modern rabies biologic products, are not risk-free, nor is their supply unlimited. The first descriptive study of PEP cases associated with the mid-Atlantic raccoon rabies epizootic during 1982-83 (133 patients) also documented that most PEP cases were due to nonbite exposures; however, these principally involved direct exposure to the suspect rabid animal (1). A 1980-81 nationwide survey of 5,634 PEP cases found an increased risk for occupational and recreational exposure to animals in a rural setting (9). The absolute mean annual PEP rate described in our report of 43 per 100,000 was nearly 10-fold higher than the rate of 4.7 per 100,000 reported in that study. A rate of 66 per 100,000 was reported from two counties (93 people per square kilometer) in New Jersey at the raccoon rabies epizootic front in 1990 (10,11). The incidence of human rabies PEP in New Jersey and this study exceeded by 10- to 20-fold the rates in areas reporting rabies in skunks (12,13), raccoons (14), and mixed wildlife (9,15). The disparity may be partially explained by regional epizootic versus enzootic status of wildlife rabies and subsequent variations in the comparative intensity of disease in wildlife populations, as well as recent increases in both human and animal population densities and their close association in suburban settings (2). The previous PEP studies involved communities in which rabies had been enzootic in terrestrial wildlife for decades (9,12-16). However, the mid-Atlantic raccoon rabies epizootic comprises the emergence of a terrestrial rabies variant into areas that had, for the most part, been free of terrestrial rabies. The exceptions were sporadic cases of spillover from geographically widespread, but low-level, bat rabies into terrestrial animals and occasional incursions of red fox rabies from Canada into New York, Vermont, and other northern states (2,16). Previous studies of PEP trends in the United States identified bites from dogs and cats as the most common animal encounter, accounting for 65% to 84% of PEP cases (7,9,12-15). By contrast, only 23% of PEP cases in this study were associated with dog or cat bites. In view of current epidemiologic trends in canine rabies-free areas of the United States, if a biting dog appears clinically normal and can be confined and observed for signs of rabies, the decision to administer PEP may be based on suggestive clinical signs and a prompt diagnostic evaluation that confirms rabies rather than on presumptively initiating PEP. Given that cats are now the leading rabid domestic animal in the United States (17), and more specifically that 12 of the 13 domestic species confirmed rabid from the four-county study area during 1993 and 1994 were cats, rabies vaccinations for cats should become more prevalent. Among exposures to owned domestic pets that resulted in human PEP, 9% of cats (versus 76% of dogs) were vaccinated against rabies. Moreover, most of the encounters with dogs that precipitated PEP in urban counties involved bites from stray dogs, indicating the need for enhanced programs for urban dog control. The economic impact of a new terrestrial rabies variant is substantial (2,18). In 1994, the New York State Department of Health increased its reimbursement to local health units for mandated rabies control activities from $75,000 to $1,080,000 to assist in the expense associated with human rabies PEP, animal rabies testing (11,896 specimens in 1993), and pet immunization clinics (114 in 1993) (4). Local health units in New York State provide funds for treatment expenses not covered by health insurance. With the cost of rabies biologic products alone exceeding $1,500 per treatment series, an exponential increase in the incidence of PEP, as documented in this study, taxes the public health infrastructure. Moreover, unlike red fox rabies, which periodically reinvades northeastern New York from adjoining areas of Canada and Vermont but then dies out, raccoon rabies is expected to persist in affected areas of New York State, as it has in the southeastern United States for the past 5 decades and in the mid-Atlantic and northeastern states more recently (17). Control of canine rabies in the United States and other industrialized countries was achieved by eliminating the susceptible reservoir population (through stray dog control and mandatory vaccination) (16). Applications of this concept to wildlife is problematic because of the difficulty in capturing wild animals for vaccination or for applying lethal measures. Population reduction alone is not sufficient to control or eliminate terrestrial wildlife rabies variants over large geographic areas (16). An emerging alternative is oral rabies vaccination of free-ranging reservoir populations, although current methods are still in their infancy and the cost-benefit of such interventions warrants further investigation (10). During the enzootic raccoon rabies in the southeastern states since the early 1950s or the current mid-Atlantic/northeastern United States epizootic, this variant has not been known to cause human rabies deaths. Yet its potential lethality for humans is supported by ample spillover into other wild animal species (predominantly skunks, but also red foxes, bobcats, and woodchucks) and into domestic animals (predominantly cats, but also dogs, cows, horses, goats, and rabbits). Substantial amounts of infectious rabies virus have been identified in the salivary glands of rabid raccoons (19). No biologic or epidemiologic data suggest unique attenuation or change in virulence of this particular rabies variant that would account for a lack of identified human deaths. Instead, epidemiologic data regarding PEP after suspected exposure to raccoon rabies indicate that PEP frequently is administered even when no exposure has been identified. Also, a bite, scratch, or other exposure, such as gross contamination of an open wound or mucous membrane with moist, infectious material from a small carnivore such as the raccoon, would unlikely be unrecognized or ignored. The apparent liberal administration of effective PEP after known bites, scratches, and other suspected exposures from rabid raccoons may have resulted in complete prevention of human deaths due to this variant of rabies virus associated with raccoons. Although the Healthy People 2000 goal to reduce PEP is worthwhile, better understanding of the circumstances leading to human exposure and formulating ways to reduce exposure is required to meet this objective. Until then, it will be particularly difficult to reduce PEP during an ever-expanding raccoon rabies epizootic. --------------------------------------------------------------------------- Acknowledgments We thank the following public health nurses and rabies coordinators for data collection and their generosity and enthusiasm in sharing data: Terri Hogan, Lynn Crane, Linda Thompson, and Hope Albino. Dr. Wyatt is chair of the University of Rochester School of Medicine and Dentistry's Division of Laboratory Animal Medicine and chief veterinarian at the Seneca Park Zoo in Rochester, New York. His research interests include the epidemiology of zoonoses in the environment and the workplace. Address for correspondence: Jeff Wyatt, University of Rochester, Box 674, Rochester, New York 14642, USA; fax: 716-273-1085; e-mail: jeff_wyatt@urmc.rochester.edu. References 1. Jenkins SR, Winkler WG. Descriptive epidemiology from an epizootic of raccoon rabies in the mid-Atlantic states 1982-1983. Am J Epidemiol 1987;126:429-37. 2. Rupprecht CE, Smith JE, Fekadu M, Childs JE. The ascension of wildlife rabies: a cause for public concern or intervention? Emerg Infect Dis 1995;1:107-14. 3. Centers for Disease Control and Prevention. Raccoon rabies epizootic—United States, 1993. MMWR Morb Mortal Wkly Rep 1994;43:269-73. 4. Hanlon CA, Trimarchi C, Harris-Valente K, Debbie JG. Raccoon rabies in New York State: epizootiology, economics and control. In: Proceedings of the 5th Annual International Meeting. Rabies in the Americas, coping with invading rabies epizootics. Niagara Falls, Canada; 1994 Nov 16-19; p. 16. 5. Public Health Service. Healthy People 2000: national health promotion and disease prevention objectives. Washington: U.S. Department of Health & Human Services; 1991. DHHS publication no. (PHS) 91-50213. 6. U.S. Department of Commerce, Economics and Statistics Administration, Bureau of the Census, 1990. Census of population, general population characteristics. New York: The Department; 1990. CP-1-34. 7. Centers for Disease Control and Prevention. Rabies prevention—United States, 1991. Recommendations of the Immunization Practices Advisory Committee. MMWR Morb Mortal Wkly Rep 1991;40:R-3:1-19. 8. Ashfar A. A review of non-bite transmission of rabies virus infection. British Veterinary Journal 1979;135:142-8. 9. Helmick CG. The epidemiology of human rabies postexposure prophylaxis, 1980-1981. JAMA. 1983;250:1990-6. 10. Uhaa IJ, Dato VM, Sorhage FE. Benefits and costs of an orally absorbed vaccine to control rabies in raccoons. J Am Vet Med Assoc 1992;201:1873-82. 11. Spencer LM. Taking a bite out of rabies. J Am Vet Med Assoc 1994;204:479-84. 12. Schnurrenburger PR, Martin RJ, Meerdink GL, Rose NJ. Epidemiology of human exposure to rabid animals in Illinois. Public Health Rep 1969;84:1078-84. 13. Martin RJ, Schnurrenberger PR, Rose NJ. Epidemiology of rabies vaccinations of persons in Illinois, 1967-68. Public Health Rep 1969;84:1069-77. 14. Currier RW, McCroan JE, Dreesen DW, Winkler WG, Parker RL. Epidemiology of antirabies treatment in Georgia, 1967-71. Public Health Rep 1975;90:435-9. 15. Winkler WG, Kappus KD. Human antirabies treatment in the United States, 1972. Public Health Rep 1979;94:166-71. 16. Baer GM. The natural history of rabies. 2nd ed. Boston: CRC Press; 1991. 17. Krebs JW, Strine TW, Smith JS, Noah DL, Rupprecht CE, Childs JE. Rabies surveillance in the United States during 1995. J Am Vet Med Assoc 1996;209:2031-44. 18. Noah DL, Smith MG, Gotthardt JC, Krebs JW, Green D, Childs JE. Mass human exposure to rabies in New Hampshire: exposures, treatment, and cost. Am J Public Health 1996;96:1149-51. 19. Winkler WG, Shaddock JS, Bowman C. Rabies virus in salivary glands of raccoons (Procyon lotor). J Wildl Dis 1985;21:297-8. _________________________________________________________________________ Dispatches _________________________________________________________________________ Dispatches Factory Outbreak of Escherichia coli O157:H7 Infection in Japan Yoshiyuki Watanabe,* Kotaro Ozasa,* Jonathan H. Mermin,† Patricia M. Griffin,† Kazushige Masuda,‡ Shinsaku Imashuku,§ and Tadashi Sawada* *Kyoto Prefectural University of Medicine, Kyoto, Japan; †Centers for Disease Control and Prevention, Atlanta, Georgia, USA; ‡Public Health Bureau, Kyoto City Government, Kyoto, Japan; and §Kyoto City Institute of Health and Environmental Sciences, Kyoto, Japan --------------------------------------------------------------------------- To determine the cause of a July 1996 outbreak of Escherichia coli O157:H7 among factory workers in Kyoto, Japan, we conducted cohort and case-control studies. Eating radish sprout salad during lunch at the factory cafeteria had been linked to illness. The sprouts were traced to four growers in Japan; one had been associated with an outbreak of E. coli O157:H7 among 6,000 schoolchildren in Sakai earlier in July. During May through August 1996, approximately 10,000 cases of Escherichia coli O157:H7 infection associated with at least 14 separate clusters were reported in Japan (1,2). Most cases occurred in school-age children. One cluster was a large outbreak in Sakai City, Osaka Prefecture, involving more than 6,000 primary school children. The outbreak started on July 13, 1996, and an investigation suggested that radish sprouts were the most likely cause (2,3). An outbreak also occurred in a factory in Kyoto, approximately 50 km from Sakai City. On July 17, 1996, a 24-year-old male factory worker went to a local clinic with diarrhea. The next day, a second worker came to the clinic with diarrhea. Bloody diarrhea and hemolytic uremic syndrome (HUS) subsequently developed in both patients, and stool cultures from each yielded E. coli O157:H7. On July 21, a third worker died of HUS-associated encephalopathy; his stool culture later yielded E. coli O157:H7. All three workers had recently eaten meals at the factory cafeteria. To identify a possible food vehicle, we conducted an epidemiologic investigation. The Study On July 19, factory officials requested that ill workers report to factory health-care workers any symptoms from the beginning of July. Stool samples from workers with diarrhea were cultured for E. coli O157:H7 and other bacterial pathogens (e.g., Salmonella and Shigella). Surveillance continued until the end of July. A culture-confirmed case was defined as a stool culture yielding E. coli O157:H7 from a factory worker who had onset of diarrhea during July 15 to 22, 1996. A clinical case was defined as diarrhea with one or more loose stools per day with onset during July 15 to 22, 1996. During their shifts, workers could eat any of the meals served at the factory cafeteria, which was operated by an outside company. Data on the date, time, and type of meal purchased at the cafeteria were routinely recorded by computer, and the cost of meals was deducted from employees' salaries. Workers could not pay by cash. We analyzed these data for visits to the cafeteria during July 8 to 14 (2 to 8 days before the date of symptom onset for the first case of culture-confirmed infection). All factory workers were included in analyses implicating a specific date of eating at the cafeteria. Only factory workers who purchased food at the cafeteria on a particular day were included in analyses of a particular meal for that day. Two set lunches with prespecified food items were served for the same price in the cafeteria each day. Because the lunches could not be distinguished by computer records, a self-administered questionnaire was completed during September 24 to 27, 1996, by the 47 workers who had reported diarrhea and 300 randomly selected workers who had eaten at the factory cafeteria on the suspected exposure days (July 11 or 12) and had not reported diarrhea in July. A computer record of the meals purchased by each of the workers was included with each questionnaire to assist with recall. Published methods were used to calculate odds ratios (ORs), 95% confidence intervals (CIs), and p values (4). P values were calculated by a chi-square test: p values of <0.05 were considered significant, and those of 0.05 to 0.09 were considered borderline significant. Multivariate conditional logistic regression analysis was conducted with Statistical Analysis System (SAS) software (SAS Institute, Cary, North Carolina, USA, 1990). After reports of the first three cases, fecal samples from ill factory workers were cultured in sorbitol indole pyruvic acid bile salts agar (SIB) medium at 35° C to 37° C for 18 to 24 hours at the Kyoto City Institute of Health and Environmental Sciences. To differentiate E. coli O157:H7 from other bacteria, colonies were examined on triple sugar iron agar, sulfide indole motility medium, lysin indole motility semisolid agar, Voges-Proskauer semisolid medium, and Simon's citrate agar. Cultures that conformed to the biochemical pattern of E. coli O157:H7 were then serotyped. The presence of Shiga toxin 1 or 2 was confirmed by reversed passive latex agglutination and polymerase chain reaction (PCR). Stored food samples were homogenized, and a portion was cultured in modified E. coli broth before culturing in SIB medium. To differentiate strains of E. coli O157:H7, pulsed-field gel electrophoresis (PFGE) and random amplified polymorphic DNA-PCR (RAPD) assays were performed as previously described (5,6). On July 18, the regional public health center examined the factory cafeteria kitchen facilities for deficiencies. All 25 food handlers were asked questions regarding abdominal symptoms and provided stool samples for bacteriologic testing. All food served was traced to the distributor and grower as far back as possible. Findings Of the 3,155 employees of the factory, 74 reported gastrointestinal symptoms in July; stool samples were obtained from these workers. Illness in 47 persons met the case definition: 42 cases were clinically defined, and 5 were culture-confirmed. The peak date of symptom onset was July 17 (Figure). Six workers had only abdominal pain, fever, or general fatigue, and 21 had onset of diarrhea outside the defined period. HUS developed in three workers with culture-confirmed E. coli O157:H7 infection; two fully recovered; one died. One clinical case-patient and four culture-confirmed case-patients had bloody diarrhea. The proportion of cases with bloody diarrhea was 11% among all patients. The median age of case-patients was 30 years (18 to 61). Of the 47 case-patients, 45 (96%) (including all culture-confirmed cases) had eaten at the factory cafeteria during July 8 to 14. Of the 47 case-patients, 39 (83%) were male, and eight (17%) were female. No information on sex and age of the other factory workers was available. Because the five [fig] workers with culture-confirmed Figure. Escherichia coli O157:H7 infection by date of E. coli O157:H7 symptom onset, July 15-21, 1996. infection had no common eating exposure except the factory cafeteria, we first analyzed the association between illness and date of eating at the cafeteria. Eating in the cafeteria any day during July 8 to 13 was associated with illness by univariate analysis. On multivariate logistic regression analysis, this association was significant or borderline significant only for July 11, 12, and 13. The ORs (95% CI, p value) of eating on July 11, 12, and 13 were 2.58 (0.91 to 7.36, 0.08), 2.84 (1.02 to 7.94, 0.05), and 3.19 (1.03 to 9.86, 0.04), respectively. Because 81% of the patients ate in the cafeteria on July 11 and 12 compared with 23% on July 13, July 11 and 12 were considered the most likely days of exposure. On multivariate analysis of the six meal times on July 11 and 12, only eating lunch in the cafeteria on July 11 was associated with illness. The rate of diarrhea for 1,134 workers who ate lunch on July 11 was 3.0%, compared with 0.6% for 2,021 workers who did not (OR = 3.04, 95% CI = 1.08, p = 0.04). Of the 47 case-patients, 44 (94%) responded to the questionnaire. In 31 patients who answered the question regarding symptoms, diarrhea lasted for a median of 3 days (1 to 10 days), and four (13%) reported bloody diarrhea. Of 300 potential controls randomly selected from factory workers who ate in the cafeteria on July 11 or 12, 291 (97%) responded to the questionnaire. Among the respondents, 16 (5%) reported gastrointestinal symptoms in July (of these, four [25%] had diarrhea during July 15-22, but none reported bloody diarrhea), and three did not respond to the question regarding symptoms; 272 respondents who reported no illness were adopted as controls. The median age was similar for case-patients (30 years [18 to 61 years]) and controls (32 years [20 to 65 years]). Eighty percent of case-patients and 83% of controls were male. Among the participants, 29 patients and 164 controls responded that they clearly remembered if they had eaten the radish sprout salad. Seventeen (59%) of 29 patients and 64 (39%) of 164 controls reported eating radish sprout salad (OR = 2.21, 95% CI = 0.99 to 4.94, p = 0.08) (Table). No other food item served on July 11 or 12 was eaten by >50% of patients and had a higher odds ratio than radish sprout salad. Among the five patients with culture-confirmed infection, computer records indicated that four patients, including the one who died, ate radish sprouts. Radish sprout salad (consisting of radish sprouts, mayonnaise, cauliflower, and fish paste) was served with both lunches the cafeteria served on July 11, the only time sprouts were served during July 8 to 14. Table. Factory cafeteria foods associated with illness, July 11, 1996, Kyoto, Japan -------------------------------------------------------------------------- Case-patients Controls exposed/total exposed/total Odds ratio(sup a) Food (%) (%) (95% CI) p value -------------------------------------------------------------------------- Radish sprout salad 17/29(58.6) 64/164(39.0) 2.21(0.99-4.94) 0.08 Boiled beef with soy sauce 8/28(28.6) 24/152(15.8) 2.13(0.84-5.40) 0.18 Scrambled eggs 10/28(35.7) 31/150(20.7) 2.11(0.89-5.04) 0.18 -------------------------------------------------------------------------- (sup a)Odds ratio>2.00; CI: confidence interval. All five patient isolates of E. coli O157:H7 produced Shiga toxins 1 and 2. E. coli O157:H7 was not detected in any of the frozen food samples (including radish sprout salad) leftover from cafeteria meals during July 11 to 15. Both PFGE and RAPD patterns of the E. coli O157:H7 isolates from this outbreak and the outbreak in Sakai City during the same time were indistinguishable (1-3,7). Examination of the factory cafeteria kitchen facilities on July 18 by the regional public health center found no deficiencies. One female food handler had diarrhea with onset July 17, but E. coli O157:H7 was not cultured from her stool specimen or from specimens of any of the other food handlers. The radish sprouts served at the cafeteria on July 11 were supplied by a single distributor that received the sprouts from four growers, one of whom also supplied the radish sprouts suspected as the source of E. coli O157:H7 infections in the Sakai City school outbreak. Radish sprouts used at the primary schools in Sakai City and at the factory cafeteria had been shipped by the grower on July 9 (3); however, the sprouts used at the factory cafeteria had been purchased along with radish sprouts from different growers. Conclusions Our data indicate that the outbreak of E. coli O157:H7 infection among Kyoto factory workers was most likely caused by contaminated radish sprouts: the factory outbreak began during the week following the Sakai City outbreak; the factory used radish sprouts from the same grower; they were shipped on the same day as those served to school children in the Sakai City outbreak; and isolates from both outbreaks had indistinguishable PFGE and RAPD patterns (1-3,5,7). The PFGE patterns of earlier outbreaks in Okayama Prefecture (Oku-cho), Gifu Prefecture, Hiroshima Prefecture, Aichi Prefecture, and Okayama Prefecture (Niimi City) were indistinguishable from each other and different from the PFGE patterns of isolates from the outbreaks in Sakai City and the Kyoto factory (1-3,5). E. coli O157:H7 was not isolated from radish sprouts; however, the process of freezing sprouts or pooling them with other food items may have decreased the number of organisms to an undetectable level. Although radish sprouts had never been linked to E. coli O157:H7 infection, they are plausible vehicles. Most outbreaks of E. coli O157:H7 infections have been linked to ground beef (8), but other items, including unrefrigerated sandwiches (9), apple cider (10), mayonnaise (11), cantaloupe (12), lettuce (13), and alfalfa sprouts (14,15) have been implicated. In addition, some sprout types, including alfalfa sprouts (16) and mung bean sprouts (17), have been linked to Salmonella outbreaks. In 1997, E. coli O157:H7 was isolated from radish sprouts collected from two different outbreaks of E. coli O157:H7 infections in Japan (1). Three cases of HUS (6%) among 47 cases of clinically or laboratory-defined cases of E. coli O157:H7 infection in the factory outbreak is comparable to rates described in other outbreaks (18,19). The proportion of workers reporting bloody diarrhea was low, possibly because infection with E. coli O157:H7 follows a more benign course in adults than in children (20) or because the amount of bacterial contamination was low. Several reasons might explain the small proportion of workers who ate lunch on July 11 and reported illness. First, some ill workers might not have informed the factory health-care personnel about gastrointestinal symptoms for fear of decreasing their chance for future promotion. This seems plausible because 16 (5%) of 291 potential controls in the case-control study mentioned unreported gastrointestinal symptoms. If this percentage of underreporting occurred for the 3,155 workers in the factory, an additional 173 infections may have been missed. Second, the pathogens might have been diluted because only a part of the radish sprout shipment was contaminated. The latter hypothesis is supported by the fact that four growers, including the one implicated in the Sakai City outbreak, supplied the radish sprouts eaten at the factory cafeteria on July 11. Third, the contamination of radish sprouts may have been reduced by washing. Although the association between eating radish sprout salad and illness among workers who ate lunch on July 11 was of borderline significance, it was the only item associated with illness that was consumed by more than 50% of case-patients. Moreover, the next lowest p value was 0.18, far from that of radish sprout salad. The radish sprout salad contained other food items; therefore, the individual risk for each food item could not be ascertained. However, radish sprouts and mayonnaise were the only uncooked ingredients. Although mayonnaise is a possible vehicle, no reports implicated it in other outbreaks in Japan in 1996 (1). Recall bias could have occurred in the case-control study because workers were asked about meals they had eaten 8 weeks earlier. Providing with the questionnaire a printout of food items purchased on July 11 and 12 may have assisted recall. In addition, as a result of the outbreak, the cafeteria stopped serving food on July 19 and had not resumed service at the time of the case-control study. This may have assisted respondents in remembering what food items they had eaten during the last week of dining in the cafeteria. --------------------------------------------------------------------------- Acknowledgments The authors thank the workers and staff of the Kyoto factory where the outbreak occurred for their cooperation in the survey; Drs. Yurie Kanamoto, Hiroyuki Imai, Shouji Tateishi, Mitsuaki Nishibuchi, Jiro Imanishi, Jun Fukuda, Yusaku Matsui, Yoichi Mori, and Kozo Yokota, and collaborators in the Kyoto City Institute of Health and Environmental Sciences; Mr. Akio Kuroda and Ms. Hisae Takenobu; and Drs. Robert Hyams and Robert Tauxe for helpful discussions and for facilitating the investigation. Dr. Watanabe is professor, Department of Social Medicine and Cultural Sciences, Research Institute for Neurological Diseases and Geriatrics, Kyoto Prefectural University of Medicine. His research focuses on the epidemiology of gastrointestinal and neurologic diseases and geriatrics. Address for correspondence: Yoshiyuki Watanabe, Department of Social Medicine and Cultural Sciences, Research Institute for Neurological Diseases and Geriatrics, Kyoto Prefectural University of Medicine, Kawaramachi-Hirokoji, Kamigyo-ku, Kyoto 602-8566, Japan; fax: 81-75-251-5770; e-mail: watanabe@basic.kpu-m.ac.jp. References 1. National Institute of Health and Infectious Diseases Control Division, Ministry of Health and Welfare of Japan. Verocytotoxin-producing Escherichia coli (entero-hemorrhagic E. coli) infections, Japan, 1996-June 1997. Infectious Agents Surveillance Report 1997;18:153-4. 2. National Institute of Health and Infectious Diseases Control Division, Ministry of Health and Welfare of Japan. Enterohemorrhagic Escherichia coli (verocytotoxin-producing E. coli) infection, 1996-April 1998. Infectious Agents Surveillance Report 1998;19:122-3. 3. Study report on the cause of the outbreak of diarrhea due to E. coli O157:H7 among primary school students in Sakai City. Tokyo: Ministry of Health and Welfare in Japan, Environmental Health Bureau, Food Sanitation Division; 1996. (In Japanese). 4. Rothman KJ. Modern epidemiology. Boston: Little, Brown and Company; 1986. p. 153-76. 5. Watanabe H, Wada A, Inagaki Y, Itoh K, Tamura K. Outbreaks of enterohaemorrhagic Escherichia coli O157:H7 infection by two different genotype strains in Japan, 1996. Lancet 1996;348:831-2. 6. Izumiya H, Terajima J, Wada A, Inagaki Y, Itoh K, Tamura K, et al. Molecular typing of Escherichia coli O157:H7 isolates in Japan by using pulsed-field gel electrophoresis. J Clin Microbiol 1997;35:1675-80. 7. Report on the cause of the outbreak of E. coli O157:H7 in Kyoto City in 1996. Kyoto, Japan: Public Health Bureau, Kyoto City Government; 1997. (In Japanese). 8. Boyce TG, Swerdlow DL, Griffin PM. E. coli O157:H7 and the hemolytic-uremic syndrome. N Engl J Med 1995;333:364-8. 9. Carter AO, Borczyk AA, Carlson JA, Harvey B, Hockin JC, Karmali MA, et al. A severe outbreak of Escherichia coli O157:H7-associated hemorrhagic colitis in a nursing home. N Engl J Med 1987;317:1496-500. 10. Besser RE, Lett SM, Weber JT, Doyle MP, Barrett TJ, Wells JG, et al. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157:H7 in fresh-pressed apple cider. JAMA 1993;269:2217-20. 11. Keene WE, Mcanulty JM, Williams LP, Hoesly FC, Hedberg K, Fleming DW, et al. A two-restaurant outbreak of Escherichia coli O157:H7 enteritis associated with consumption of mayonnaise [abstract]. In: Proceedings of the 33rd Interscience Conference on Antimicrobial Agents and Chemotherapy; 1993 Oct 17-20; New Orleans, Louisiana. Washington: American Society for Microbiology; 1993. p. 354. 12. Yet another outbreak of hemorrhagic colitis—Corvallis. Portland (OR): Center for Disease Prevention and Epidemiology, Oregon Health Division, Department of Human Resources. CD Summary 1993;42:1-2. 13. Mermin JH, Hilborn ED, Voetsch A, Swartz M, Lambert-Fair MA, Farrar J, et al. A multistate outbreak of Escherichia coli O157:H7 infections associated with eating mesclum mix lettuce [abstract]. In: Proceedings of the 3rd International Symposium and Workshop on Shiga Toxin (verocytotoxin)-producing Escherichia coli infections; 1997 Jun 22-26; Baltimore, Maryland. The Lois Joy Galler Foundation; 1997. p. 9. 14. Outbreaks of Escherichia coli O157:H7 infection associated with eating alfalfa sprouts—Michigan and Virginia, Jun-Jul 1997. MMWR Morb Mortal Wkly Rep 1997;46:741-4. 15. Mohle-Boetani JC, Werner SB, Farrar JA, Abbott S, Bryant R, Vugia DJ. Outbreak of toxigenic E. coli O157:H7: non-motile (NM) associated with a clover-alfalfa mix [abstract]. In: Program and Abstracts of the Infectious Diseases Society of America 36th Annual Meeting; 1998 Nov 12-15; Denver, Colorado. 536 Fr, 178. 16. Mahon BE, Ponka A, Hall WN, Komatsu K, Dietrich SE, Siitonen A, et al. An international outbreak of Salmonella infections caused by alfalfa sprouts grown from contaminated seeds. J Infect Dis 1997;175:876-82. 17. O'Mahoney M, Cowden J, Smyth B, Lynch D, Hall M, Rowe B, et al. An outbreak of Salmonella saint-paul infections associated with bean sprouts. Epidemiol Infect 1990;104:229-35. 18. Bell BP, Goldoft M, Griffin PM, Davis MA, Gordon DC, Tarr PI, et al. A multistate outbreak of Escherichia coli O157:H7associated bloody diarrhea and hemolytic uremic syndrome from hamburgers: the Washington experience. JAMA 1994;272:1349-53. 19. Akashi S, Joh K, Tsuji A, Ito K, Hoshi H, Hayakawa T, et al. A severe outbreak of haemorrhagic colitis and haemolytic uraemic syndrome associated with Escherichia coli O157:H7 in Japan. Eur J Pediatr 1994;153:650-5. 20. Rodrigue DC, Mast EE, Greene KD, Davis JP, Hutchinson MA, Wells JG, et al. A university outbreak of Escherichia coli O157:H7 infection associated with roast beef and an unusually benign clinical course. J Infect Dis 1995;172:1122-5. -------------------------------------------------------------------------- Dispatches First Case of Yellow Fever in French Guiana since 1902 J.M. Heraud,* D. Hommel,† A. Hulin,† V. Deubel,‡ J.D. Poveda,‡ J.L. Sarthou,* and A. Talarmin* *Institut Pasteur de la Guyane, Cayenne, French Guiana;†General Hospital, Cayenne, French Guiana; and ‡Institut Pasteur, Paris, France --------------------------------------------------------------------------- The first case of yellow fever in French Guiana since 1902 was reported in March 1998. The yellow fever virus genome was detected in postmortem liver biopsies by seminested polymerase chain reaction. Sequence analysis showed that this strain was most closely related to strains from Brazil and Ecuador. Yellow fever (YF) is a serious public health problem in many tropical countries in Africa and South America. In South America, most infections are sporadic, affecting unvaccinated persons who enter the forest; monkeys are the primary reservoirs, and Haemagoggus sp. are the vectors. French Guiana, located between Brazil and Surinam in the Amazonian forest, has had many epidemics of YF. However, no case has been reported in French Guiana since 1902, although a serologic survey in 1951 found circulating virus; of 430 persons younger than 50 years of age (who had not been affected by the 1902 epidemic), 9% of those living in the coastal area and 29% of those living inland had significant titers of neutralizing YF antibodies (1). A study conducted at the same time in Surinam showed even higher rates of YF seropositivity in persons who had not been exposed to previous epidemics and confirmed that YF viruses were circulating in the region (2). In French Guiana, YF immunization became compulsory in 1967. Case Report In March 1998, an Amerindian woman living in a forest area on the Maroni River was admitted to the health center in Maripasoula, French Guiana, with fever, headache, abdominal pain, vomiting, and diarrhea. She was treated for malaria because of a Plasmodium falciparum–positive blood smear. Two days later, the patient's fever increased (40.2°C), she became jaundiced, and she was evacuated to the intensive care unit (ICU) at Cayenne Hospital with multiple visceral failure: shock syndrome, renal failure (blood urea level 32 mmol/l, creatinine level 656 µmol/l), and liver failure (total bilirubin level 314 µmol/l, alanine aminotransferase 2048 IU/l, aspartate amino transferase 6256 IU/l, prothrombin level 23%). No hemorrhages were noted. Despite symptomatic treatment, the condition of the patient deteriorated rapidly, and she died a few hours after admission to ICU. Blood cultures were negative for bacterial pathogens. Because of anuria, urine cultures were not possible, and albuminuria could not be tested. Examination of peripheral blood smears showed no parasites on admission to ICU and titers of antibodies to leptospira were low. Microscopy examination of postmortem liver biopsies showed histopathologic changes characteristic of YF: midzonal necrosis with a small rim of a few viable periportal and pericentral hepatocytes and centrilobular cells with microsteatosis and eosinophilic degeneration with round, eosinophilic cytoplasmic structures (Councilman bodies). A serum sample collected before death and a serum sample obtained from the patient in 1994 during a seroepidemiologic study on HTLV-I infection and stored at -80°C at the Institut Pasteur de la Guyane, French Guiana, were tested for immunoglobulin (Ig) G and IgM specific for three flaviviruses (YF, dengue, and Saint Louis encephalitis), two alphaviruses Tonate and Mayaro), and a new world arenavirus (Tacaribe), by enzyme-linked immunosorbent assay (3). A plaque reduction neutralization test was also used to detect YF-neutralizing antibodies (4). The serum collected in 1998 contained IgM (but not IgG) that specifically recognized YF antigens. IgM specific for other flavivirus antigens (dengue, Saint Louis encephalitis) were not found. The neutralizing antibody titer of the 1998 serum as assessed by plaque reduction neutralization test was 20. In the serum collected in 1994, no YF virus–specific antibodies were detected by any technique. IgG to Mayaro virus was detected in both sera; no other antibodies to alphaviruses and arenaviruses were detected. Serum and homogenized liver samples from the patient were diluted 10-fold in Leibowitz medium containing 3% fetal calf serum, and dilutions were injected into subconfluent AP 61 cell cultures (5). After 7 days, cells were harvested and tested for YF virus by an indirect immunofluorescence assay using a monoclonal antibody from the Centers for Disease Control and Prevention, Fort Collins, CO, USA. YF virus was not isolated by cell culture from either blood or liver samples. Reverse transcription–polymerase chain reaction (RT-PCR) tests were performed on RNA extracted from serum and liver (6). RNA of the YF virus was detected by RT-PCR after seminested PCR only in the liver sample. The 542-bp PCR product was purified and directly sequenced with an automatic sequencing system (ACTgene, EuroSequence Gene Services, Evry, France). The first 309 nucleotides of the 3' noncoding sequence were aligned with those of sequenced YF strains from Genbank (7,8). Sequences were aligned with the Clustal W program. Bootstrap confidence limits were calculated from 100 replicates with the program SEQBOOT. Phylogenetic analyses were performed by maximum parsimony by using the DNAPARS program with uniform character weights and a heuristic search option. All branch lengths were drawn to scale by the program Treetool. The sequence of the YF strain isolated in 1998 was deposited in Genbank (accession number: AF121952). The sequence of the amplified gene differed at 11 positions (3.5% nucleotide divergence) from that of a Brazilian strain isolated in 1935 and at 21 positions (6.8% nucleotide divergence) from that of a strain isolated in Ecuador in 1979. The sequence diverged more from strains isolated in Peru (1995) and Trinidad (1979) (8.1% and 10% divergence, respectively). As expected, African strains differed more, with those of West Africa (from Nigeria and Senegal) being less distant than those from East and Central Africa (Uganda and Central African Republic). The nucleotide sequence downstream from the NS5 stop codon in the 3' noncoding region of the YF virus was deleted in the French Guianese strain, as in several South American YF viruses (9). Phylogenetic analysis of the sequences confirmed that the virus was most closely related to those isolated in Brazil and Ecuador (Figure). Conclusions [Fig] The histopathologic Figure. Phylogenetic tree generated from 309 changes of the nucleotides of the 3' noncoding region of the strain liver were of yellow fever (YF) isolated in French Guiana in 1998 characteristic of (in bold) and of 14 other YF strains by using the YF but also of DNAPARS program. Numbers indicate bootstrap values for other hemorrhagic groups to the right. One µl (30 ng) of primer VD8 fever viruses. The (5'-GGGTCTCCTCTAACCTCTAG-3') was mixed with RNA IgMs were only resuspended in 10 µl of distilled water; the mixture slightly above the was heated at 95°C for 2 minutes and placed on ice. cut-off values cDNA was synthesized in a mixture containing reverse used in the transcriptase (RT) incubation buffer (provided by the laboratory. manufacturer), 0.2 mM of each of the four dNTPs, 20 Although YF was units RNasine, and five units of AMV RT (Promega, highly probable, Charbonnières, France) by incubation for 1 hour at the diagnosis 42°C. cDNA was amplified by PCR. Four µl of the cDNA required sample was added to 46 µl of a mixture containing Taq confirmation by polymerase buffer (provided by the manufacturer), 2 mM detection of the MgCl2, 0.5 mM of each of the 4 dNTPs, 300 ng of primer virus or its VD8 and of degenerate primer EMF1 genome. Indeed, in (5'-TGGATGACSACKGARGAYATG-3') (S = C, G; K = G, T; R = 1990 a suspected A, G; Y = C, T), and 0.5 unit of Taq polymerase case of YF was (Promega, Charbonnières, France). After 5 minutes of reported to the denaturation at 95°C, the mixture was subjected to 30 World Health polymerase chain reaction (PCR) cycles: 95°C for 30 Organization seconds, 53°C for 90 seconds, and 72°C for 60 seconds, because of followed by a final 10-minute polymerization step at characteristic 72°C. Four µl of a 1 in 100 dilution of the PCR histopathologic products was used for seminested PCR using primers VD8 changes, but the and NS5YF (5'-ATGCAGGACAAGACAATGGT-3'). After the case was never denaturation step, DNA was amplified by 25 cycles of confirmed PCR: 94°C for 30 seconds, 55°C for 90 seconds, and (presence of IgG 72°C for 120 seconds, followed by a final extension but no IgM step at 72°C. Negative controls (serum from healthy specific for YF, persons) were included in the series. The positive no detection of control (supernatant of infected mosquito cells) was the virus in liver tested separately to avoid any contamination. The or serum); later phylogenetic analysis was conducted by the Pasteur it was shown that Institute in Paris. the patient had been vaccinated against YF 1 year before (10). However, because YF requires health authorities to take specific measures, confirmation of the diagnosis is important, especially when YF is not prevalent. The case we described was confirmed by RT-PCR from liver only, because the serum sample was taken on day 6 when viremia is usually resolved and because the sensitivity of cell culture for virus in liver samples is very low (probably because of biliary salts toxicity). This case underscores the need for postmortem liver biopsies for detecting the viral genome to confirm diagnosis. This patient did not leave her village the months before infection; she was probably infected while working in forest clearings. The detection of a YF virus in French Guiana nearly a century after the last report is notable; however, the absence of reported cases during the previous years is surprising because no natural borders exist between this country and northern Brazil, where YF is not uncommon (11). A severe YF outbreak would have easily been detected, but sporadic cases can be misdiagnosed as other fevers or as hepatitis (when jaundice is present) and may be not tested for YF even though serologic and YF virus detection tests are performed for each suspected case. No other case was diagnosed in the patient's family or neighborhood, but sporadic cases are common in South America (12), probably because of poorly anthropophilic vectors. Furthermore, in this area, approximately 90% of the population have been vaccinated at least once (R. Pignoux, unpub. data). Outbreaks are common among nonhuman primates, but no epidemic has occurred in the area where the patient lived. However, YF incidence increased in northern Brazil in 1998 (13). This case calls attention to vaccination problems in French Guiana, especially along the rivers. Our patient had been vaccinated in 1985, but the absence of neutralizing antibodies in 1994 indicates that the vaccination was not effective. Although this patient may have had a poor antibody response, more likely inadequate storage of the vaccine before use was responsible. In 1985, YF vaccines were less thermostable than now, and the cold chain was difficult to maintain. In response to this case, an immunization campaign was initiated in this area in May. Vaccination of the population must continue since YF can reappear. The immunization program implemented in French Guiana (compulsory vaccination of children older than 1 year of age, booster YF vaccination every 10 years, and required vaccination certificate before entering school) should avert the threat of outbreaks in urban areas, which have a vaccine coverage rate of approximately 80%. However, the risk for sporadic cases in unvaccinated persons will persist, and so active serologic and virologic surveillance of YF remains necessary. --------------------------------------------------------------------------- Acknowledgment The authors thank Michel Favre for his help in the phylogenetic analysis of the yellow fever virus sequences. This study was supported by a grant from the Programme Hospitalier de Recherche Clinique. Jean-Michel Heraud is a Ph.D. student working in the virology laboratory of the Pasteur Institute of French Guiana. His areas of expertise are virology and hematology with a focus on arbovirus epidemiology and physiopathology. Address for correspondence: A. Talarmin, Institut Pasteur de la Guyane, 23 Avenue Pasteur, BP 6010, 97306 Cayenne Cedex, Guyane Française; fax: 0594-309-416; e-mail: atalarmin@pasteur-cayenne.fr. References 1. Floch H, Durieux C, Koerber R. Enquête épidémiologique sur la fièvre jaune en Guyane française. Annales Institut Pasteur 1953;84:495-508. 2. Wolff JW, Collier WA, De Roever-Bonnet H, Hoekstra J. Yellow fever immunity in rural population groups of Surinam. Tropical and Geographical Medicine 1958;325-31. 3. Lhuillier M, Sarthou JL. Intérêt des IgM anti-amariles dans le diagnostic et la surveillance épidémiologique de la fièvre jaune. Annales de Virologie (Institut Pasteur) 1983;134E:349-59. 4. Lindsey HS, Calisher CH, Matthews JH. Serum dilution neutralization test for California group virus identification and serology. J Clin Microbiol 1976;4:503-10. 5. Reynes JM, Laurent A, Deubel V, Telliam E, Moreau JP. The first epidemic of dengue hemorrhagic fever in French Guiana. Am J Trop Med Hyg 1994;51:545-53. 6. Deubel V, Huerre M, Cathomas G, Drouet M-T, Wuscher N, Le Guenno B, et al. Molecular detection and characterization of yellow fever in blood and liver specimens of a non-vaccinated fatal human case. J Med Virol 1997;53:212-7. 7. Hahn CS, Dalrymphe JH, Strauss JH, Rice CM. Comparison of the virulent Asibi strain of yellow fever virus with the 17D vaccine strain derived from it. Proc Natl Acad Sci U S A 1987;84:2019-23. 8. Wang E, Ryman KD, Jenings AD, Wood DJ, Taffs F, Minor PD, et al. Comparison of the genomes of the wild-type French viscerotropic strain of yellow fever virus with its vaccine derivative French neurotropic vaccine. J Gen Virol 1995;76:2749-55. 9. Wang E, Weaver SC, Shope RE, Tesh RB, Watts DM, Barett ADT. Genetic variation in yellow fever virus: duplication in the 3' noncoding region of strains from Africa. Virology 1996;225:274-81. 10. World Health Organization. Yellow fever in 1989 and 1990. Wkly Epidemiol Rec 1992;67:245-51. 11. World Health Organization. Yellow fever in 1994 and 1995. Wkly Epidemiol Rec 1996;71:313-8. 12. Tolou H. La fièvre jaune: aspects modernes d'une maladie ancienne. Méd Trop 1996;56:327-32. 13. World Health Organization. Yellow fever in Brazil. Wkly Epidemiol Rec 1998;7:351. -------------------------------------------------------------------------- Dispatches Risk for Rabies Transmission from Encounters with Bats, Colorado, 1977–1996 W. John Pape,* Thomas D. Fitzsimmons,*† and Richard E. Hoffman* *Colorado Department of Public Health and Environment, Denver, Colorado, USA; and †Centers for Disease Control and Prevention, Atlanta, Georgia, USA --------------------------------------------------------------------------- To assess the risk for rabies transmission to humans by bats, we analyzed the prevalence of rabies in bats that encountered humans from 1977 to 1996 and characterized the bat-human encounters. Rabies was diagnosed in 685 (15%) of 4,470 bats tested. The prevalence of rabies in bats that bit humans was 2.1 times higher than in bats that did not bite humans. At least a third of the encounters were preventable. Although no cases of human rabies have been reported since 1931 in Colorado, rabies remains a health risk in this state because of the frequency with which Coloradans have contact with bats. The first objective of this study was to determine the prevalence of rabies in bats that were submitted for laboratory testing in Colorado over a 20-year period, including an analysis by bat species. The second objective was to characterize the circumstances of confirmed bat-human encounters during this same period and to evaluate how this information could be used to prevent human rabies. Data Sources Laboratory Records Rabies diagnosis was conducted by two laboratories in Colorado: the Colorado Department of Public Health and Environment (CDPHE) Laboratory and the Colorado State University (CSU) Veterinary Diagnostic Laboratory. Bats were accepted for testing from public and private sources if they had had contact with a person or a domestic pet, if the possibility of contact could not be excluded, or if the bat exhibited abnormal behavior. County agencies were also permitted to submit up to three bats per week (usually found dead from no apparent cause or exhibiting aberrant behavior) for local surveillance. None of the submissions were for studies of rabies prevalence among bats with a normal appearance or in their natural habitat. Records from both laboratories were maintained by the CDPHE Epidemiology Division and made up the first dataset we analyzed. Information extracted from these records included rabies test date, test result, and bat bite information. For bats sent to CDPHE (but not CSU), laboratory technicians identified the bats by species, and the data were included in the analysis. Possible Rabies Exposure Memoranda A second dataset consisted of memoranda describing any animal exposure reported to CDPHE resulting in rabies postexposure prophylaxis (PEP). Memoranda were not written for encounters in which the animal tested negative for rabies, even if a person was bitten. Infrequently, CDPHE staff wrote memoranda before learning that an animal had tested negative for rabies or when PEP was recommended but not administered. Four persons in the Epidemiology Division worked on zoonosis control during the 20-year study period; two of them wrote 90% of the memoranda. A bat encounter was defined as bat contact or possibility of bat contact with a person. A wound was defined as a visible puncture, scratch, bleeding, or a sensation of sharp pain during the encounter. No attempt was made to distinguish bite wounds from claw marks or scratches. The analysis of circumstances was restricted to memoranda in which the presence of a bat was documented. Memoranda that described encounters with bats were identified, and data on person, place, and time were extracted. The circumstance of encounter was listed as one of 13 general categories that best described the event. Any person who initiated the standard rabies PEP series was designated as having received treatment. Findings Laboratory Records From 1977 through 1996, 4,502 bats were submitted for testing. CDPHE received 4,394 bats (98%), and CSU received 108 bats (2%). These bats represented 15 (83%) of 18 species present in Colorado (1) (Table 1). Thirty-two bats were excluded from further analysis because either the test result or the bite status was unrecorded. Rabies was diagnosed in 685 (15%) bats and accounted for 98% of all animal rabies cases in Colorado during the study period. Of the 233 bats that bit people, 69 (30%) had rabies. Of the 4,237 bats that did not bite people, 613 (14%) had rabies. The prevalence of rabies among bats that bit humans was 2.1 times higher (95% confidence interval 1.7 to 2.5) than in bats not involved in human bites. None of the persons bitten by bats got rabies. Human rabies has not been reported in Colorado since 1931 (CDPHE, unpub. data, 1998). Table 1. Prevalence of rabies in bats submitted for Species testing, Colorado, 1977–1996 data were ------------------------------------------------------------------- available No. for No. that did 4,182 that bit not bite Total no. (94%) humans humans tested bats. Species (% rabid) (% rabid) (% rabid) Three ------------------------------------------------------------------- species—big brown Big brown bat 122 2,013 2,135 bats Eptesicus fuscus (27) (16) (17) (Eptesicus Myotis genus group(sup a) 35 722 757 fuscus), (14) (6) (7) hoary bats Silver-haired bat 28 628 656 (Lasiurus cinereus), Lasionycteris (14) (5) (5) and noctivagans silver-haired Hoary bat 13 452 465 bats (Lasionycteris Lasiurus cinereus (77) (39) (40) noctivagans)—accounted Long-eared bat 8 38 46 for Myotis evotis (88) (21) (33) 73% of total Brazilian free- 0 41 41 submissions, tailed bat 84% of (0) (12) (12) the Tadarida rabies-positive brasiliensis specimens, and Red bat 1 25 26 70% of Lasiurus borealis (100) (8) (12) bats involved Pallid bat 0 21 21 in Antrozous pallidus (0) (5) (5) bite Big free-tailed bat 2 19 21 incidents. The Nyctinomops (50) (11) (14) prevalence macrotis of Townsend's big- 1 13 14 rabies eared bat (0) (0) (0) in silver-haired Plecotus townsendii bats Species data 23 265 288 (5%) was unavailable (35) (9) (11) lower Total 233 4,237 4,470 than (30) (14) (15) in big brown ------------------------------------------------------------------- bats (sup a)Includes six species in the genus Myotis that could not be (17%) easily distinguished by inspection: M. lucifugus, M. volans, M. or thysanodes, M. californicus, M. ciliolabrum, and M. yumanesis. hoary bats(40%). Memoranda During the 20-year study period, 271 memoranda described possible encounters with bats; 240 (89%) memoranda documented the presence of a bat and were included in the analysis. Of the 131 bats tested, 99 had rabies. Of the 240 persons who encountered bats, 141 (59%) were male and 99 (41%) were female. From the 195 (81%) records that recorded the person's age, the range in age was 10 months to 81 years, and the median age was 25 years. Of the 182 (76%) persons reporting that they were wounded, the most common wound site was the hand (59%), followed by the arm (14%), head/neck (12%), leg/foot (9%), torso (2%), or multiple sites (2%). Not enough information was available to characterize the time of day of the encounters. Two hundred (83%) encounters occurred between June and September, corresponding to peak activity periods and seasonal migratory patterns of bats in Colorado. In the 217 records that noted location of bat encounters, 117 (54%) occurred outdoors, chiefly on home properties and park and recreation areas (none were reported in caves). Of the 100 (46%) bat encounters inside buildings, 83 were in private homes (37 of these in bedrooms). Big brown bats, colonial bats that commonly roost in houses and buildings, were encountered more frequently outside than inside (60% vs. 39%) when rabid (n = 46). All 11 rabid hoary bats, solitary tree dwellers, were encountered outdoors. However, rabid silver-haired bats, another solitary tree-roosting species, were encountered equally indoors (n = 3) and outdoors (n = 3). The four most frequent circumstances in which people encountered bats, accounting for 62% of the encounters, were a bat landing on an awake person (19%), a person picking up a grounded bat outside (18%), a person awakening to find a bat in the room (15%), and a person trying to remove a bat from inside a structure (10%). The remaining nine circumstances occurred repeatedly but less frequently (Table 2). Of the 240 persons who had encounters with bats, 216 (90%) initiated PEP; nine of these stopped treatment after the bat tested negative for rabies. The bat tested negative in 17 of the 24 cases in which PEP was not administered, but a memorandum was written before the test results were available. The remaining seven persons did not receive prophylaxis because they or their physician did not believe that the contact warranted treatment. In three of these seven encounters (which occurred before 1983), the bat was found to be rabid, but no definite wound was observed. The time from bat Table 2. Circumstances in which humans encountered encounter to bats, Colorado, 1977–1996 initiation of treatment could be ------------------------------------------------------- calculated for 199 Bat captured (92%) of the 216 and tested Bat All patients who Not not encoun received PEP and Circumstances Rabid rabid tested -ters was 1 hour to 28 days. Fifty percent ------------------------------------------------------- of patients Bat landed on person 17 2 27 46 received their Person picked up 24 5 15 44 first dose of bat outdoors vaccine within 24 hours of exposure; Person awoke to 17 4 14 35 75% started find bat in room treatment within 72 Person tried to remove 5 2 17 24 hours. Of the 18 bat from indoors patients who initiated treatment Person inadvertently 3 5 8 16 7 or more days touched hidden bat after the Person handled 12 0 1 13 encounter, nine did captured bat not do so until Child found alone 4 3 2 9 advised by an with bat acquaintance or physician of the Person handled bat 6 1 1 8 possible rabies as part of job risk. Person stepped on bat 3 0 3 6 Person bitten while 2 1 3 6 Silver-Haired Bats taking bat from pet Although the Person bitten by pet that 1 4 1 6 silver-haired bat had bat in mouth rabies virus Person attributed wound 0 2 0 2 variant was to bat they saw isolated from 15 of the 21 persons who Other circumstances 0 0 6 6 died of Unspecified in report 5 3 11 19 bat-associated Total 99 32 109 240 rabies in the United States from ------------------------------------------------------- 1980 through 1997, we observed that silver-haired bats in Colorado had neither the greatest frequency nor the highest species-specific rate of rabies. Our findings are consistent with tabulations from New York (1988 to 1992) and Arkansas, Virginia, and West Virginia (1990 to 1994), which showed that silver-haired bats made up a small proportion of bats submitted for rabies testing; only a small number of submitted silver-haired bats were rabies positive (2,3). Nonetheless, Arkansas (1991), New York (1993), and West Virginia (1994) each had human cases associated with the rabies virus variant common to silver-haired bats (4-6). Because the frequency of human encounters with this species is apparently low and the prevalence of rabies in tested silver-haired bats is small, other factors must explain the silver-haired bats' association with human rabies cases. One hypothesis is that silver-haired bats are more aggressive than other bats (1). Additionally, one study has demonstrated that the rabies virus variant of silver-haired bats replicates in nonneuronal tissue more efficiently than a coyote rabies virus variant (7). This attribute might explain how a small dermal inoculum of silver-haired variant rabies virus from a seemingly superficial bite could cause infection. As silver-haired and hoary bats are tree dwellers that favor old growth forest habitat, it should be unexpected to encounter them indoors. None of the 12 hoary bats (11 rabid) included in this series were encountered inside. In contrast, three of the nine encounters with silver-haired bats were indoors. All three bats were rabid. Conclusions Bats that interact with humans are far more likely to have rabies than bats that avoid humans, and rabies prevalence is highest in bats that bite. Conversely, rabid bats appear to interact more frequently and to be more prone to bite than nonrabid bats. This behavior is consistent with clinical manifestations of rabies in wildlife species in which the animal exhibits abnormal behavior, loses its natural fear of humans, and acts aggressively (8). Encounters with bats resulted from a relatively small number of recurring situations. At least a third of the encounters (picking up grounded bats, handling captured bats, and trying to remove bats from structures or from pets' mouths) were preventable; however, most encounters in which a person inadvertently touched a hidden bat or a bat landed on a person were probably unavoidable. Delays in treatment suggest that some people may not be aware of the risk for rabies transmission from contact with a bat. In the United States, in nearly half of the cases of human rabies associated with bat variant rabies virus, the person had no history of contact with a bat (9-11). Although unrecognized exposures may have occurred, the persons involved probably did not understand the risk after exposure to a bat and therefore did not seek medical care. Even when specifically asked about animal exposures, some patients and their families initially did not report bat contact (11,12). This study has several potential limitations. First, we do not know whether the prevalence of rabies in tested bats is representative of all bats that encounter humans. The laboratory testing was a passive surveillance system, dependent on the submission of bats by persons involved in encounters. We calculated a lower limit estimate of the prevalence of rabies among bats that bit humans by using data from the memoranda. If one assumed that the bats that bit humans and later escaped and thus were not tested (90) were rabies-free, the prevalence of rabies among bats that bit people would decrease from 30% (69 of 233) to 21% (69 of 323), still significantly higher than the prevalence in bats that did not bite people (p < .001). The circumstances were categorized for a small proportion of all bat encounters. Memoranda were written for only 131 of 4,502 bats submitted for testing and for 109 encounters in which the bat escaped. The number of unreported encounters cannot be estimated, and information from those encounters could alter the frequencies of the type of encounter presented in this study. Finally, administration of rabies PEP is not reportable in Colorado and, therefore, the study may not have included all persons who received PEP after a bat encounter. Because the state health department was the primary source of rabies biological supplies for medical providers in the state from 1977 through 1985, nearly all exposures requiring PEP would have come to the department's attention. Rabies biological supplies were more widely available from other sources after 1985. Although the number of reported PEP administrations remained stable, some exposures may not have been reported. Two findings of this study support recent revisions of the Advisory Committee on Immunization Practices (ACIP) recommendations (13-15): bats that encountered humans had a high prevalence of rabies, and the third most frequently reported circumstance was a person awakening to find a bat in the room. The ACIP stated in October 1997 that PEP may be appropriate even in the absence of demonstrable bite, scratch, or mucous membrane exposures in situations in which such exposure is likely to have occurred (e.g., a sleeping person awakes to find a bat in the room or an adult finds a bat in a room with an unattended child, a mentally deficient person, or an intoxicated person) (14). Of 35 instances reported in this study in which a bat was found in the room by a person upon awakening, 17 bats were rabid, and 23 persons had evidence of a bite. The Colorado Department of Public Health and Environment, the Colorado Division of Wildlife, and the Colorado Bat Society recently collaborated to publish an educational pamphlet that describes methods to prevent rabies exposure from a bat and measures to take if a person encounters a bat (16). --------------------------------------------------------------------------- Acknowledgments We thank rabies laboratory workers for invaluable service over the past 20 years, especially Larry Briggs and Jane Carman. We also thank John Emerson for his diligence in investigating and documenting bat encounters from 1977 to 1986. John Pape is an epidemiologist with the Communicable Disease Program, Colorado Department of Public Health and Environment. He has statewide responsibility for zoonotic disease surveillance and control. In addition to bat rabies, his research focuses on the epidemiology of zoonotic diseases in the western United States, including plague, hantavirus, and tick-borne relapsing fever. Address for correspondence: W. John Pape, 4300 Cherry Creek S, Denver, CO 80246, USA; fax: 303-782-0904; e-mail: john.pape@state.co.us. References 1. Armstrong DM, Adams RA, Navo KW, Freeman J, Bissell SJ. Bats of Colorado: shadows in the night. 2nd ed. Denver (CO): Colorado Division of Wildlife; 1996. p. 11. 2. Childs JE, Trimarchi CV, Krebs JW. The epidemiology of bat rabies in New York state, 1988-92. Epidemiol Infect 1994;113:501-11. 3. Dreesen DW, Orciari LA, Rupprecht CE. The epidemiology of bat rabies in the southeastern United States 1990-1994 [abstract]. In: Proceedings from 7th Annual International Meeting of Advances Towards Rabies Control in the Americas; 1996 Dec 9-13; Atlanta, Georgia. p. 44. 4. Centers for Disease Control. Human rabies—Texas, Arkansas, and Georgia, 1991. MMWR Morb Mortal Wkly Rep 1991;40:765-9. 5. Centers for Disease Control and Prevention. Human rabies—New York, 1993. MMWR Morb Mortal Wkly Rep 1993;42:799,805-6. 6. Centers for Disease Control and Prevention. Human rabies—West Virginia, 1994. MMWR Morb Mortal Wkly Rep 1995;44:86-7,93. 7. Morimoto K, Patel M, Corisdeo S, Hooper DC, Fu ZF, Rupprecht CE, et al. Characterization of a unique variant of bat rabies responsible for newly emerging human cases in North America. Proc Natl Acad Sci U S A 1996;93:5653-8. 8. Kaplan C, Turner GS, Warrell DA. Rabies: the facts. 2nd ed. Oxford: Oxford University Press; 1986. p. 72-4. 9. Centers for Disease Control and Prevention. Human rabies—Montana and Washington, 1997. MMWR Morb Mortal Wkly Rep 1997;46:770-4. 10. Centers for Disease Control and Prevention. Human rabies—Texas and New Jersey, 1997. MMWR Morb Mortal Wkly Rep 1997;47:1-5. 11. Centers for Disease Control. Human rabies—Texas, 1990. MMWR Morb Mortal Wkly Rep 1991;40:132-3. 12. Centers for Disease Control and Prevention. Human rabies—Connecticut, 1995. MMWR Morb Mortal Wkly Rep 1996;45:207-9. 13. Advisory Committee on Immunization Practices. Revised ACIP rabies post-exposure prophylaxis (PEP) statement1997 Oct 22. 14. Constantine DG. Bat rabies in the southwestern United States. Public Health Rep 1967;82:867-8. 15. Rabies prevention—United States, 1991: recommendations of the Immunization Practices Advisory Committee (ACIP). MMWR Morb Mortal Wkly Rep 1991;40:1-19. 16. Colorado Division of Wildlife, Colorado Department of Public Health and Environment, Colorado Bat Society. Bats and rabies [pamphlet]. Denver (CO): The Department; 1997. -------------------------------------------------------------------------- Dispatches Australian Bat Lyssavirus Infection in a Captive Juvenile Black Flying Fox Hume Field,* Brad McCall,† Janine Barrett‡ *Queensland Department of Primary Industries, Moorooka, Australia; †Brisbane Southside Public Health Unit, Upper Mount Gravatt, Australia; and ‡The University of Queensland, Brisbane, Australia --------------------------------------------------------------------------- The newly emerging Australian bat lyssavirus causes rabieslike disease in bats and humans. A captive juvenile black flying fox exhibited progressive neurologic signs, including sudden aggression, vocalization, dysphagia, and paresis over 9 days and then died. At necropsy, lyssavirus infection was diagnosed by fluorescent antibody test, immunoperoxidase staining, polymerase chain reaction, and virus isolation. Eight human contacts received postexposure vaccination. Australia was considered free of rabies and the rabieslike viruses of the genus Lyssavirus until the recognition in 1996 of Australian bat lyssavirus (ABL) as the cause of a rabieslike disease in a black flying fox (Pteropus alecto)(1) and a wildlife caretaker (2). While serotypic, antigenic, and sequence analysis show that ABL is closely related to classic rabies virus and European bat lyssavirus (1), phylogenetic analysis has clearly demonstrated that ABL represents a new genotype, genotype 7 (3). Australia is still considered free of terrestrial (genotype 1) classic rabies (4). Rabies vaccine and antirabies immunoglobulin protect laboratory animals against ABL infection (5), and their use pre- and post-ABL exposure is recommended for humans (6,7). On the morning of December 8, 1997, two juvenile black flying foxes were found clinging to each other and vocalizing in a residential area near an urban flying fox colony. An experienced, rabies-vaccinated wildlife caretaker retrieved the two animals from an unusually low tree roost. On the basis of body weight and forearm measurements, their age was estimated at 2 to 3 weeks, an age of total maternal dependency. Their physical condition was normal. Both animals, Bat 1 (male) and Bat 2 (female), remained with the original caretaker for 2 days before being placed with two different caretakers for hand-rearing. Bat 1 was communally housed with another orphaned black flying fox, and Bat 2 was housed with two others. For the next 5 weeks, all the bats were clinically normal. However, in week 6, Bat 1 began to exhibit signs of neurologic disease. The caretaker first observed the bat's sudden and progressive aggression toward its companion and separated them. Throughout day 1 of illness, the bat periodically "frothed at the mouth" and had repeated lordotic spasms, during which it vocalized loudly. Treatment with oral amoxycillin was initiated. On day 2, the bat was calmer but still vocal, attempting to bite objects and eating little. On day 3, it was no longer aggressive and was only able to eat pulped food and milk. On day 4, it was seen by a veterinarian, who noted severe pharyngitis, and administered injectable dexamethasone. The bat was much more alert that evening and ate solid food well. The dexamethasone injection was repeated on day 5; the bat remained alert and ate solid food overnight. On day 6, it was dysphagic and was again offered pulped foods and liquids. On days 7 and 8, it was unable to roost normally, lay supine, was progressively dysphagic, had diarrhea, and was losing weight. On day 9, it rapidly got worse and died. The carcass was submitted to the Queensland Department of Primary Industries Animal Research Institute for necropsy. The emaciated carcass had poor pectoral muscle development and no perirenal, pericardial, or mesenteric fat reserves. ABL infection was diagnosed by fluorescein-labeled antirabies monoclonal globulin (CENTOCOR) in a fluorescence antibody test (FAT) on impression smears of fresh brain. With the absence of other lyssaviruses in Australia (1), a positive reaction with this lyssavirus genus-specific antibody is considered diagnostic for ABL. Sections of brain showed nonspecific, nonsuppurative meningoencephalitis with perivascular cuffs of mononuclear cells and widespread focal gliosis. Numerous neurons contained eosinophilic inclusion bodies, which are highly suggestive of lyssavirus infection. Immunoperoxidase staining of formalin-fixed, paraffin-embedded sections of brain with a monoclonal antinucleoprotein antibody (Clone HAM, provided by R. Zanoni, Berne University, Switzerland) detected lyssavirus antigen in neurons of the frontal cortex, hippocampus, brain stem, and cerebellum, including Purkinje cells. This antigen distribution is consistent with previous reports of rabies (8). The diagnosis was independently confirmed at the Commonwealth Scientific and Industrial Research Organisation's Australian Animal Health Laboratory by FAT, immunoperoxidase staining, virus isolation in murine neuroblastoma cells, and sequence analysis of polymerase chain reaction (PCR) product (P. Daniels, pers. comm.). No blood was available for serologic testing. After ABL infection was diagnosed in Bat 1, the Brisbane Southside Public Health Unit received information that up to eight persons had been bitten or scratched by the bat in the weeks before and during its illness. Six were bat handlers who had received postexposure treatment with five doses of rabies human diploid cell vaccine (HDCV) during a 1996 campaign that followed the diagnosis of the first human case of ABL (2,7). Two were unvaccinated members of the principal bat caretaker's household. Despite recommendations that unvaccinated members of bat caretakers' households not handle bats, the two had come into regular contact with the bat and may have been scratched during that time. Lyssavirus prophylaxis was commenced in accordance with Australian recommendations (6,7). All eight persons provided blood for rabies serologic testing (indirect-enzyme linked immunosorbent assay [ELISA]). The six vaccinated bat handlers had titers of 1.130 IU/ml to >8.80 IU/ml 12 months after their initial vaccinations (World Health Organization_recommended protective level for rabies = 0.5 IU/ml [9]). Each of these received two further intramuscular doses of 1.0 ml of HDCV, according to the Australian recommendations for postexposure treatment of vaccinated persons. The two unvaccinated persons had titers of <0.13 IU/ml (nonimmune) and received the standard postexposure treatment for unvaccinated persons: 20 IU per kg of human rabies immune globulin (HRIG) by intramuscular injection and five intramuscular doses of 1.0 ml of HDCV. All eight remain well 10 months after the incident. Throughout this episode, Bat 2 and the other three bats that had been directly or indirectly in contact with the infected bat remained healthy. After the diagnosis of ABL in Bat 1, these four animals were quarantined for observation at the Animal Research Institute for 11 weeks and then euthanized. All were antibody-negative (<0.5 IU/ml) by rabies rapid fluorescent focus inhibition test when quarantined and remained so during the observation period. Brain impression smears from the four were negative for lyssavirus antigen by FAT. This case of naturally occurring ABL infection is of particular interest for several reasons. First, the astute observations of the bat caretaker provide possibly the first record of the clinical course of natural ABL infection in a flying fox. Second, to our knowledge, this is the first recorded case of ABL disease in a maternally dependent juvenile. Third, the case history provides the first indication of the incubation period and length of clinical disease in naturally infected flying foxes. Natural in utero infection with lyssaviruses is not known to occur (10), and the four flying foxes in contact with Bat 1 were unlikely to be the source of infection as they subsequently tested negative for ABL antibody and antigen. The infection appears to have occurred in the 2 to 3 weeks before the rescue and, after an incubation period of 6 to 9 weeks, produced 9 days of clinical disease. Infection in Bat 1 most probably resulted from a bite from an ABL-infected bat. This bat may have been Bat 1's dam. This episode demonstrates the necessity of examining for ABL any flying fox that has bitten or scratched a person and of improving community and professional awareness of the disease and associated risks. Costly postexposure treatment with HRIG can be avoided if only vaccinated persons handle Australian bats. --------------------------------------------------------------------------- Acknowledgments We thank Helen Luckhoff, Jill Nelson, Jennifer Hawyes, and Merle Thomas for reporting the case and making their records available; Barry Rodwell for performing the fluorescence antibody test; Natasha Smith and Craig Smith for technical assistance; Greg Smith, Ina Serafin, and Alan Westacott for performing ELISA (human) serology; Ross Lunt, Peter Hooper, and Alan Gould for fluorescence antibody test and immunoperoxidase staining, rapid fluorescent focus inhibition test (flying fox) serology, virus isolation, and sequence analysis of PCR product; and the general practitioners and the staff of Queen Elizabeth II Hospital for collecting serum samples and administering postexposure treatments. Hume Field is a veterinary research officer with the Animal Research Institute, Queensland Department of Primary Industries. His particular research interest is the epidemiology of wildlife diseases and the role of wild species as reservoirs of infection for domestic animals and humans. He is investigating the natural history of Australian bat lyssavirus and equine morbillivirus (Hendra virus), two recently emerged zoonotic diseases in Australia. Address for correspondence: Janine Barrett, School of Veterinary Science and Animal Production, The University of Queensland, Brisbane, Queensland, Australia 4072; fax: 61-7-3362-9420; e-mail: j.barrett@mailbox.uq.edu.au. References 1. Fraser GC, Hooper PT, Lunt RA, Gould AR, Gleeson LJ, Hyatt AD, et al. Encephalitis caused by a lyssavirus in fruit bats in Australia. Emerg Infect Dis 1996;2:327-31. 2. Allworth A, Murray K, Morgan J. A human case of encephalitis due to a lyssavirus recently identified in fruit bats. Commun Dis Intell 1996;20:504. 3. Gould AR, Hyatt AD, Lunt R, Kattenbelt JA, Hengstberger S, Blacksell SD. Characterisation of a novel lyssavirus isolated from Pteropid bats in Australia. Virus Res 1998;54:165-87. 4. Office International des Epizooties, 10 Nov 1998, [computer program]. Handistatus (download). Version 1.39. Microsoft Internet Explorer 4.0. http://www.oie.int/A_html.htm. 5. Hooper PT, Lunt RA, Gould AR, Samaratunga H, Hyatt AD, Gleeson LJ, et al. A new lyssavirus—the first endemic rabies-related virus recognised in Australia. Bulletin Institut Pasteur 1997;95:209-18. 6. Rabies and bat lyssavirus infection. In: Watson C, editor. The Australian Immunisation Handbook. 6th ed. Canberra: Australian Government Publishing Service; 1997. p. 162-8. 7. Prevention of human lyssavirus infection. Commun Dis Intell 1996;20:505-7. 8. Feiden W, Kaiser E, Gerhard L, Dahme E, Gylstorff B, Wandeler A, et al. Immunohistochemical staining of rabies virus antigen with monoclonal and polyclonal antibodies in paraffin tissue sections. Zentralbl Veterinarmed [B] 1988;35:247-55. 9. World Health Organization recommendations on rabies post-exposure treatment and the correct technique of intradermal immunization against rabies. Geneva: The Organization; 1997. 10. Constantine DG. Absence of prenatal infection of bats with rabies virus. J Wildl Dis 1986;22:249-50. -------------------------------------------------------------------------- Dispatches Bordetella holmesii-Like Organisms Isolated from Massachusetts Patients with Pertussis-Like Symptoms W. Katherine Yih, Ellen A. Silva, James Ida, Nancy Harrington, Susan M. Lett, and Harvey George Massachusetts Department of Public Health, Boston, Massachusetts, USA --------------------------------------------------------------------------- We isolated Bordetella holmesii, generally associated with septicemia in patients with underlying conditions, from nasopharyngeal specimens of otherwise healthy young persons with a cough. The proportion of B. holmesii-positive specimens submitted to the Massachusetts State Laboratory Institute increased from 1995 to 1998. Bordetella holmesii is a recently described gram-negative, asaccharolytic, nonoxidizing, soluble, brown-pigment-producing rod previously known as CDC nonoxidizer group 2 (NO-2) (1). This group consists of 15 closely related, biochemically similar strains of fastidious nonmotile bacteria isolated from human blood cultures. In establishing NO-2 as a species, Weyant et al. (1) performed 16S rRNA sequencing of one NO-2 strain and the type strains of B. pertussis, B. parapertussis, B. bronchiseptica, and B. avium. They found a high degree of homology among them (>98% over 1,525 bases) and confirmed a close relatedness between NO-2 and Bordetella species by DNA relatedness studies (hydroxyapatite method). Biochemically, the lack of oxidase activity and the production of a brown soluble pigment differentiate B. holmesii from B. pertussis, B. bronchiseptica, and B. avium; the lack of urease activity differentiates it from B. parapertussis (1). Unlike B. pertussis, which causes whooping cough, B. holmesii has been associated most often with septicemia in patients with underlying conditions (1-4). It also has been isolated from sputum from one patient with respiratory symptoms (3). Van den Akker (5) suggested that the difference in lipopolysaccharide expression (important in bacterial pathogenesis) between the closely related B. pertussis and B. holmesii might help explain their observed difference in propensity to infect respiratory tract epithelium verses causing opportunistic bacteremia. In Massachusetts, however, we have seen B. holmesii associated with a different clinical picture. From January 1995 (when the article describing B. holmesii [1] was published) through December 1998, the Massachusetts State Laboratory Institute (SLI) isolated B. holmesii from 34 clinical specimens: 33 nasopharyngeal specimens from patients suspected of having pertussis and one blood culture specimen from a 45-year-old patient with septicemia. Of the 33 patients with respiratory symptoms, 30 (91%) were 11 to 29 years old, 1 (3%) was an infant, and 2 (6%) were 10 years old; most were otherwise healthy. B. holmesii is confirmed by its biochemical patterns and cellular fatty acid analysis or the DNA transformation test will definitively separate B. holmesii from Acinetobacter, neither procedure is performed at SLI. Three of the initial isolates were sent to the Centers for Disease Control and Prevention, Atlanta, Georgia, for definitive identification and were confirmed as B. holmesii by cellular fatty acid analysis. The remaining 31 isolates were biochemically and morphologically identical to those. B. holmesii–positive nasopharyngeal specimens have increased both in absolute number and as a percentage of the total nasopharyngeal specimens processed at SLI (Table). The number rose from 3 (0.1% of total specimens submitted for pertussis culture) in 1995, to 6 (0.2%) in 1996, to 9 (0.4%) in 1997, and to 15 (0.6%) in 1998. A chi-square analysis for linear trend of proportions from 1995 to 1998 found the trend significant (chi-square = 11.6, p <.001), possibly indicating a rise in prevalence. (When such an analysis was applied to the proportion of nasopharyngeal specimens testing positive for pertussis during the same period, no trend was apparent ([chi-square = 0.8, p = .36].) A growing awareness of the species by laboratory personnel may be contributing to the observed increase. Table. Bordetella species isolated from nasopharyngeal (NP) specimens at the Massachusetts State Laboratory Institute, 1994–1998 ---------------------------------------------------------------------- 1994 1995 1996 1997 1998 -------------------------------------------------- No. % No. % No. % No. % No. % ---------------------------------------------------------------------- B. pertussis 75 4.2 140 5.8 325 8.9 132 5.6 165 6.6 B. parapertussis 7 0.4 20 0.8 32 0.9 11 0.5 NA NA B. holmesii 0 0 3(sup a)0.1 6 0.2 9 0.4 15 0.6 Total NP specimens 1,792 2,399 3,653 2,375 2,508 reported ---------------------------------------------------------------------- (sup a)Does not include the one case of B. holmesii isolated from blood. To ascertain what symptoms were associated with B. holmesii isolation, we investigated the 23 cases identified in 24 months from January 1997 through December 1998. We called providers and patients for disease histories and demographic information. Pertussis case report forms, modified to include more possible symptoms and underlying conditions, were used to record the information collected in the interviews. Nineteen (83%) of the 23 B. holmesii-positive cases were in adolescents (11 to 19 years), 2 (9%) in young adults (20 and 29 years), 1 (4%) in a 10-year-old child, and 1 (4%) in an infant. All had cough. In addition, 14 (61%) had paroxysms, 2 (9%) had whoop, and 6 (26%) had posttussive vomiting. No other symptoms were identified by patients or providers. Fourteen (61%) of the 23 had no underlying conditions, 8 (35%) had minor conditions such as occasional asthma or allergies, and 1 (4%) had chronic fatigue. The fact that cultures were taken from 20 (87%) of the 23 case-patients within 14 days of cough onset generally excluded convalescent-stage pertussis as a cause of symptoms. However, B. pertussis had been confirmed in a 14-year-old girl, who had occasional asthma, 3 months before B. holmesii was confirmed—she received a reculture because of a persistent cough that had not resolved since the original infection. She had had paroxysms and vomiting associated with the pertussis infection but no symptoms other than cough at the time B. holmesii was isolated. Cultures were taken on the same day from two sisters, 15 and 9 years old, each with cough of fewer than 14 days. B. holmesii was culture-confirmed in the 15-year-old; B. pertussis was culture-confirmed in the 9-year-old. This raises the question of whether B. pertussis and B. holmesii might cocirculate. At least 11 (48%) of the 23 cases were found during active surveillance for pertussis in school and university settings, which may explain the age profile of the cases. Three case-patients, with cough onset dates of April 1, 1997; February 27, 1998; and March 9, 1998, were students at the same university. These cases, though not epidemiologically linked to each other, were in symptomatic contacts of pertussis patients and were cultured as part of an azithromycin efficacy study. At least eight other cases were also in contacts of confirmed patients with pertussis and were detected through active surveillance. No cases were definitively linked. Peak months of cough onset were November and December (8 of the 23 cases), as is true for pertussis in Massachusetts, with a smaller peak in March and April (6 of the 23 cases). The observed peak in November-December may be due to the role of active surveillance for pertussis in ascertaining cases of B. holmesii colonization. The clinical profile of the 21 cases in adolescent and adult patients infected with B. holmesii was compared with that of 122 culture-confirmed pertussis cases in patients between 11 and 29 years of age with cough onsets in the same period, i.e., 1997 through 1998. (Relevant clinical information was not available for an additional 42 culture-confirmed pertussis case-patients in this age group.) We did not consider cough duration, because of imprecise data, but rather focused on the presence or absence of three classic pertussis symptoms: paroxysms of cough, whoop, and posttussive vomiting. Cases were categorized as patients with 0, 1, or 2 to 3 of these symptoms. (No separate category for three symptoms was used due to a cell size of 0 in the case of B. holmesii.) On applying the chi-square test for independent proportions, we found B. holmesii infection milder (i.e., accompanied by fewer of the above three pertussis symptoms) than B. pertussis infection (chi-square = 10, p <.01). To rule out the possibility that the difference was due to more frequent cultures for severe than for milder pertussis cases, we compared the 21 adolescent and adult B. holmesii patients with the 577 SLI–serology-positive pertussis patients 11 to 29 years of age with cough onsets in 1997 or 1998 for whom sufficient clinical data were available. The SLI pertussis serology test is a single-serum enzyme-linked immunosorbent assay for immunoglobulin G to pertussis toxin, available since 1987. The assay is for use in persons > 11 years of age and is optimally sensitive at 2 to 8 weeks after cough onset. By the same methods as for the previous comparison, we found that the B. holmesii cases were milder at a higher level of significance (chi-square = 69, p <10(sup -8). Without knowing the prevalence of B. holmesii carriage in asymptomatic persons, we cannot say with certainty that B. holmesii is the causative agent for the respiratory symptoms of the patients from whom it was isolated. Approximately half the cases were discovered through active surveillance for pertussis. This, together with the fact that the symptoms associated with B. holmesii were relatively mild, suggests that the organism may not have been causing disease. On the other hand, B. holmesii may be the etiologic agent, given that it is closely related to B. pertussis and the associated symptoms (like those of B. parapertussis) are similar. B. pertussis is the only Bordetella species known to produce pertussis toxin, although B. parapertussis and B. bronchiseptica have silent copies of the toxin gene (6). We do not know whether B. holmesii has the toxin gene. However, since B. parapertussis can cause disease (albeit not as severe as B. pertussis [7]), the presence of pertussis toxin is not necessary for the development of symptoms. Continued investigation, including conducting diagnostic tests for agents such as Chlamydia and Mycoplasma and culturing symptomatic and asymptomatic contacts, is warranted to ascertain the degree to which B. holmesii is pathogenic in the respiratory system and contagious. If it is contagious, antibiotic susceptibility testing is also needed. B. holmesii is susceptible to some 15 antibiotics of a variety of classes (2,3), but whether erythromycin, the drug of choice for pertussis, is effective is not known. SLI is establishing routine erythromycin susceptibility testing of B. pertussis and will also test B. holmesii isolates. --------------------------------------------------------------------------- Acknowledgment We thank Colin D. Marchant for early suggestions regarding data analysis. Dr. Yih is epidemiology coordinator for vaccine-preventable diseases at the Massachusetts Department of Public Health. Her research interests include ecologic and evolutionary aspects of infectious disease. Address for correspondence: W. Katherine Yih, State Laboratory Institute, 5th floor, 305 South Street, Boston, MA 02130, USA; fax: 617-983-6840; e-mail: Katherine.Yih@state.ma.us. References 1. Weyant RS, Hollis DG, Weaver RE, Amin MFM, Steigerwalt AG, O'Connor SP, et al. Bordetella holmesii sp. nov., a new gram-negative species associated with septicemia. J Clin Microbiol 1995;33:1-7. 2. Lindquist SW, Weber DJ, Mangum ME, Hollis DG, Jordan J. Bordetella holmesii sepsis in an asplenic adolescent. Pediatr Infect Dis J 1995;14:813-5. 3. Tang Y-W, Hopkins MK, Kolbert CP, Hartley PA, Severance PJ, Persing DH. Bordetella holmesii-like organisms associated with septicemia, endocarditis, and respiratory failure. Clin Infect Dis 1998;26:389-92. 4. Morris JT, Myers M. Bacteremia due to Bordetella holmesii. Clin Infect Dis 1998;27:912-3. 5. van den Akker WMR. Lipopolysaccharide expression within the genus Bordetella: influence of temperature and phase variation. Microbiol 1998;144:1527-35. 6. Arico B, Rappuoli R. Bordetella parapertussis and Bordetella bronchiseptica contain transcriptionally silent pertussis toxin genes. J Bacteriol 1987;169:2847-53. 7. Mastrantonio P, Stefanelli P, Giuliano M, Herrera Rojas Y, Ciofi degli Atti M, Anemona A, et al. Bordetella parapertussis infection in children: epidemiology, clinical symptoms, and molecular characteristics of isolates. J Clin Microbiol 1998;36:999-1002. -------------------------------------------------------------------------- Dispatches New Cryptosporidium Genotypes in HIV-Infected Persons Norman J. Pieniazek,* Fernando J. Bornay-Llinares,* Susan B. Slemenda,* Alexandre J. da Silva,* Iaci N. S. Moura,* Michael J. Arrowood,* Oleg Ditrich,† and David G. Addiss* Centers for Disease Control and Prevention, Atlanta, Georgia, USA; and †Institute of Parasitology AS CR, Ceské Budëjovice, Czech Republic --------------------------------------------------------------------------- Using DNA sequencing and phylogenetic analysis, we identified four distinct Cryptosporidium genotypes in HIV-infected patients: genotype 1 (human), genotype 2 (bovine) Cryptosporidium parvum, a genotype identical to C. felis, and one identical to a Cryptosporidium sp. isolate from a dog. This is the first identification of human infection with the latter two genotypes. Protozoan apicomplexan parasites from the genus Cryptosporidium infect a wide variety of hosts (1). The parasites are transmitted to humans through contaminated drinking water (2), contact with infected animals, and contact with infected persons (3). In the immunocompetent, cryptosporidiosis manifests itself as self-limited diarrhea, sometimes accompanied by nausea, abdominal cramps, fever, and vomiting. In the immunodeficient, however, cryptosporidiosis may be severe, chronic, and life-threatening (4). Cross-infection experiments, in which Cryptosporidium oocysts were obtained from animals of one species and fed to animals of another species, have investigated the host specificity of this parasite (5). The differences observed in the host range of putative C. parvum isolates led to a proposal to establish the Cryptosporidium isolate originating from guinea pigs, morphologically indistinguishable from C. parvum, as a new species—C. wrairi—solely on the basis of experimental infection (6). The possibility of many Cryptosporidium species fostered the development of techniques suitable for typing isolates. Commonly used techniques are isoenzyme analysis (7), Western blotting (8,9), random amplified polymorphic DNA analysis (10-12), polymerase chain reaction restriction fragment length polymorphism (PCR-RFLP) analysis (13,14), and PCR followed by DNA sequencing (11,15-19). Early studies of the polymorphism of isolates classified as C. parvum found significant geographic variation among isolates (20) in the region coding for the small subunit ribosomal RNA (SSU-rRNA), commonly used for taxonomic classification. Recently, it has been shown (21) that one of the sequences used in this analysis (22) was erroneously identified as a C. parvum sequence, while in fact it was C. muris. More recent work (e.g., Le Blancq et al. [23] and GenBank entry AF040725) has shown that the SSU-rRNA region of the C. parvum zoonotic (bovine) genotype does not show heterogeneity and is practically identical to the sequence submitted to GenBank in 1993 (accession number L16996, [24,25]). Recently, consistent results of typing bovine and human C. parvum isolates led to unequivocal recognition of two genotypes of C. parvum. These two genotypes were reproducibly differentiated by sequencing the SSU-rRNA coding region (16), sequencing the Cryptosporidium thrombospondin-related adhesion protein (TRAP-C2) gene (17), and PCR-RFLP analysis of the Cryptosporidium oocyst wall protein (COWP) gene (15). Thus far, the anthroponotic genotype (genotype 1) has been found only in infected humans, while the zoonotic genotype (genotype 2) has been found both in infected humans and in livestock, e.g., cows, lambs, goats, and horses. The published partial SSU-rRNA sequence of the C. parvum genotype 1 (16) is identical to the Cryptosporidium SSU-rRNA sequence that we observed in PCR-amplified DNA from a patient with AIDS (GenBank accession number L16997). That the data from different laboratories for these two genotypes were reproducible calls into question the recent conclusion of Tzipori and Griffiths (26) that "there are no clearly defined and fully characterized reference `strains' of Cryptosporidium." Other new Cryptosporidium genotypes have been identified in various animals (pigs, mice, cats, and koalas) by sequencing the region coding for SSU-rRNA (16,27,28). The cat genotype is thought to represent C. felis (28). When morphologic and other data are available for these new genotypes, they may be recognized as new Cryptosporidium species. We present results of typing Cryptosporidium isolates by direct sequencing the variable region of the SSU-rRNA amplified from fecal specimens with Cryptosporidium genus-specific diagnostic PCR primers (25). Applying this method to a set of specimens from a 3-year longitudinal study on the risk for enteric parasitosis and chronic diarrhea in immunodeficient patients with a low CD4+ count, we found the first human cases of infection by C. felis and by a newly identified zoonotic Cryptosporidium species possibly originating from a dog. The Study All analyzed specimens were collected from January 1991 through September 1994 in a study assessing the impact of enteric parasite-associated diarrhea in persons infected with HIV (29). Study participants answered comprehensive questionnaires concerning clinical and epidemiologic information and provided stool specimens monthly. All stool specimens were examined for C. parvum by Kinyoun carbol-fuchsin modified acid-fast stain (30) and direct immunofluorescence (31). Stained slides were examined by observing 200 oil-immersion fields. C. parvum was associated with 18 (5.1%) of the 354 acute episodes and 36 (12.9%) of the chronic episodes of diarrhea reported by the participants. All specimens from this study were preserved in separate vials of 10% formalin and polyvinyl alcohol (Para-Pak Stool System; Meridian Diagnostics, Inc., Cincinnati, OH) and thus were not suitable for molecular analysis. Some specimens were originally aliquoted and stored without preservation at -80°C. Of this set, we selected 18 available specimens for 10 randomly chosen Cryptosporidium-positive patients for this study. An aliquot of approximately 300 µl of each stool specimen was suspended in 1 ml of 0.01 M phosphate-buffered saline, pH 7.2, containing 0.01 M of EDTA (PBS-EDTA), and the suspension was centrifuged at 14,000 x g, 4°C for 5 minutes. The pellet from this centrifugation was washed two more times under the same conditions. The pellet was resuspended in 300 µl of PBS-EDTA and used for DNA extraction, performed with the FastPrep disrupter and the FastDNA kit (BIO 101, Inc., Vista, CA) (32). Extracted DNA was stored at 4°C until PCR amplification. Cryptosporidium genus-specific primers (CPBDIAGF and CPBDIAGR) were used to amplify the Cryptosporidium SSU-rRNA variable region (24,25). The conditions for PCR were 95°C for 15 minutes; 45 cycles of denaturation at 94°C for 30 seconds, annealing at 65°C for 30 seconds, extension at 72°C for 1 minute, 30 seconds; followed by extension at 72°C for 9 minutes; and finished with a hold step at 4°C. PCR products were analyzed by electrophoresis on 2% SeaKem GTG agarose (Cat. No. 50074, FMC Bioproducts, Rockland, ME), stained with ethidium bromide, and visualized on UV transilluminator. The size of the diagnostic fragment amplified with these primers for C. parvum genotype 1 (GenBank accession number L16997) was 438 bp, for C. parvum genotype 2 (L16996) 435 bp, for C. baileyi (L19068) 428 bp, and for C. muris (L19069) 431 bp. PCR products were purified by using the Wizard PCR Preps kit (Cat. No. A7170, Promega, Madison, WI). Sequencing reactions were done with the Perkin Elmer Big Dye kit (Cat. No. 4303149, PE Biosystems, Foster City, CA) and analyzed on the Perkin Elmer ABI 377 automatic DNA sequencer. Sequences were assembled by using the program SeqMan II (DNASTAR Inc., Madison, WI). Sequences were aligned with the program MACAW (33) and analyzed by programs from the PHYLIP (phylogeny inference program) package (34). Findings After the 18 fecal specimens were retrieved from storage and processed for PCR with the CPBDIAGF/CPBDIAGR diagnostic primer set, visualization of PCR products on 2% agarose gels found two different sizes of diagnostic bands. The size of the Cryptosporidium diagnostic bands for patients 53, 75, 119, 124, 153, 278, and 554 did not differ significantly from the size of the standard band (approximately 435 bp) obtained with cloned SSU-rRNA for the C. parvum genotype 2 (Figure 1, lane 1). On the other hand, the Cryptosporidium diagnostic bands in samples from patients 84, 184, and 485 were visibly larger, approximately 450 bp (Figure 1, lane 3 shows data for patient 84). [Fig] DNA sequence analysis showed Figure 1. Agarose gel (2%) visualization four types of sequences of diagnostic polymerase chain reaction (Table 1). The sequence of products of four Cryptosporidium the diagnostic band from genotypes. Lane S, standard 100-bp ladder; patient 53 was 435 bp long lane 1, patient 53, zoonotic genotype 2; and was identical to the lane 2, patient 119, Cryptosporidium sp. corresponding fragment of the (zoonotic, canine genotype); lane 3, C. parvum genotype 2 (GenBank patient 84, C. felis (zoonotic, feline accession number L16996) genotype); lane 4, patient 75, SSU-rRNA. The diagnostic anthroponotic genotype 1. bands in patients 75, 124, 153, 278, and 554 were 438 bp long and were identical to the corresponding fragment of the C. parvum genotype 1 SSU-rRNA (GenBank accession number L16997). The sequence from patient 119 was 429 bp long; the diagnostic bands from patients 84, 184, and 485 were 455 bp long and were identical to each other. Sequence similarity searches of GenBank database and molecular phylogeny analysis of the two latter types of sequences showed that, while they were not identical to any sequences in GenBank, they clustered within other representative SSU-rRNA sequences from the genus Cryptosporidium (results not shown). In addition, these sequences were significantly different from recently reported partial SSU-rRNA sequences (16,19,27,28). The 429-bp-long diagnostic fragment in patient 119 was identical to a recently identified canine C. parvum isolate SSU-rRNA sequence (GenBank accession number AF112576). The longest sequence (455 bp) from patients 84, 184, and 495 was identical in a 100-bp overlap to an updated sequence (GenBank accession number AF097430) for C. felis (also called feline C. parvum genotype) (28). Alignment of these sequences is shown in Figure 2. In the hypervariable alignment region (from position 101 to 110), all genotypes display a variable number of thymidines. Genotype 1 has the longest stretch, 11 thymidines, while the Cryptosporidium sp. (canine genotype) has only one. The sequence for C. felis has five insertions at alignment positions 45, 86, 111, 187, and 218, as well as one deletion of two thymidines at positions 197 and 198. Apart from the difference in the hypervariable region, the Cryptosporidium sp. (canine genotype) has a deletion of two bases at positions 49 and 50. Results of genotyping the 18 specimens are in Table 2. For patients 53, 124, 153, 184, 278, and 485, only a single specimen was available. For other patients (e.g., patient 554), four specimens collected during 12 months were available. The same Cryptosporidium genotype persisted throughout a patient's infection. Table 1. Cryptosporidium genotypes for 10 selected patients in this study ----------------------------------------------------------- Patient no. Band size(sup a)(bp) C. parvum genotype ----------------------------------------------------------- 53 435 Zoonotic (genotype 2) 75, 124, 153 438 Anthroponotic 278, 554 (genotype 1) 119 429 Cryptosporidium sp. (canine genotype) 84, 184, 485 455 C. felis ----------------------------------------------------------- (sup a)Size of the Cryptosporidium diagnostic band obtained with the CPBDIAGF/CPBDIAGR PCR primer pair [Fig] Figure 2. Alignment of the Cryptosporidium small subunit ribosomal DNA (SSU-rRNA) diagnostic fragments obtained with the CPBDIAGF/CPBDIAGR polymerase chain reaction primer pair for the four genotypes. Only the first 300 columns of the alignment are shown, as the remaining columns were identical for all genotypes. Gaps are shown with dashes (-), and bases identical to the base in the first row (C. felis) are shown with dots (.). Numbers to the right of the alignment show sequence positions for each genotype. The sequences for the SSU-rRNA diagnostic fragment of all four genotypes were submitted to GenBank and were assigned accession numbers AF087574, AF087575, AF087576, and AF087577. Table 2. Persistence of Cryptosporidium genotypes in patients(sup a) ------------------------------------------------------------- Month of follow-up No. of ------------------------------------- Patient available no. specimens 1 2 3 4 5 6 7 8 9 10 11 12 ------------------------------------------------------------- 53 1 B 75 2 H - - H 84 4 F - F - F F 119 2 C - C 124 1 H 153 1 H 184 1 F 278 1 H 485 1 F 554 4 H - - - - - - H - H - H ------------------------------------------------------------- (sup a)B, C. parvum genotype 2 (zoonotic, bovine); C, Cryptosporidium sp. (zoonotic, canine); F, C. felis (zoonotic, feline); H, C. parvum genotype 1 (anthroponotic); -, specimen not available. Conclusions Using the sequence of a diagnostic fragment of SSU-rRNA, as well as two well-established genotypes of C. parvum (anthroponotic genotype 1 and zoonotic genotype 2), we detected two new Cryptosporidium genotypes. The first, in patients 84, 184, and 485, was identical to the feline Cryptosporidium genotype (28), also described as C. felis. The second, in two specimens from patient 119, represents the newly identified Cryptosporidium sp. found in a sequence originating from a dog (GenBank accession number AF112576). New taxons may be established within the genus Cryptosporidium on the basis of the isolate's host range together with molecular data, even though morphologic criteria are apparently lacking. We propose to use C. felis Iseki, 1979, instead of feline C. parvum genotype; we believe that the anthroponotic genotype 1 of C. parvum and the Cryptosporidium sp. (canine genotype) described here should be named as distinct species. Although the taxonomy of protists is morphology-based, the taxonomy of bacteria is being reevaluated on the basis of molecular data (35). There is no reason to use different approaches to these two kingdoms, nor is it necessary to postulate that the basis for speciation of Cryptosporidium remains ambiguous or that molecular data lend little support to separation of Cryptosporidium into distinct, valid species (26). Further, we see no evidence for the unconventional conclusions of Tzipori and Griffiths that "...genetic markers in C. parvum can change upon passage to a different host, possibly through a selective mechanism favoring different populations" (26) nor for those of Widmer (21) that "...genetic studies and experimental infections suggest that a selective mechanism triggered by a change in the intestinal environment might be involved in shaping the genetic make-up of C. parvum populations." The finding of new Cryptosporidium genotypes in immunodeficient patients might suggest a unique susceptibility to infections by divergent Cryptosporidium species circulating in companion animals or livestock. However, recent identification of C. felis in a cow (36) may indicate a complex pattern of flow of different Cryptosporidium species in the environment. When more data on the distribution of these species become available, public health and preventive measures will need to be reevaluated, and isolates may need to be renamed to reflect their natural history. Finally, it is clear that Kinyoun carbol-fuchsin modified acid-fast stain (30) and direct immunofluorescence (31) could detect all genotypes, but this may not be true for diagnostic reagents that may be affected directly (PCR) or indirectly (Ab) by the genetic composition of the new Cryptosporidium isolates reported here. Thus, discovery of these new Cryptosporidium genotypes in human cryptosporidiosis should cause existing diagnostic tools to be reevaluated. --------------------------------------------------------------------------- Acknowledgment The authors thank Jane L. Carter for her expert help with running the DNA sequencer. Dr. Pieniazek is a microbiologist with the National Center for Infectious Diseases, CDC. His research interests include molecular reference diagnosis of parasitic diseases and molecular systematics. Address for correspondence: Norman J. Pieniazek, Division of Parasitic Diseases, Centers for Disease Control and Prevention, 4770 Buford Highway NE, Mail Stop F13, Atlanta, GA, 30341-3724, USA; fax: 770-488-4108; e-mail: nxp3@cdc.gov. References 1. Fayer R, Speer CA, Dubey JP. The general biology of Cryptosporidium. In: Fayer R, editor. Cryptosporidium and cryptosporidiosis. Boca Raton (FL): CRC Press; 1997. p. 1-41. 2. MacKenzie WR, Schell WL, Blair KA, Addiss DG, Peterson DE, Hoxie NJ, et al. Massive outbreak of waterborne Cryptosporidium infection in Milwaukee, Wisconsin: recurrence of illness and risk of secondary transmission. Clin Infect Dis 1995;21:57-62. 3. Navin TR. Cryptosporidiosis in humans: review of recent epidemiologic studies. Eur J Epidemiol 1985;1:77-83. 4. Arrowood MJ. Diagnosis. In: Fayer R, editor. Cryptosporidium and cryptosporidiosis. Boca Raton (FL): CRC Press; 1997. p. 43-64. 5. Lindsay DS. Laboratory models of cryptosporidiosis. In: Fayer R, editor. Cryptosporidium and cryptosporidiosis. Boca Raton (FL): CRC Press; 1997. p. 209-23. 6. Vetterling JM, Jervis HR, Merrill TG, Sprinz H. Cryptosporidium wrairi sp. n. from the guinea pig Cavia porcellus, with an emendation of the genus. Journal of Protozoology 1971;18:243-7. 7. Awad-el-Kariem FM, Robinson HA, Dyson DA, Evans D, Wright S, Fox MT, et al. Differentiation between human and animal strains of Cryptosporidium parvum using isoenzyme typing. Parasitology 1995;110:129-32. 8. McLauchlin J, Casemore DP, Moran S, Patel S. The epidemiology of cryptosporidiosis: application of experimental sub-typing and antibody detection systems to the investigation of water-borne outbreaks. Folia Parasitol 1998;45:83-92. 9. Nichols GL, McLauchlin J, Samuel D. A technique for typing Cryptosporidium isolates. Journal of Protozoology 1991;38:237S-40. 10. Carraway M, Widmer G, Tzipori S. Genetic markers differentiate C. parvum isolates. J Eukaryot Microbiol 1994;41:26S. 11. Morgan UM, Constantine CC, O'Donoghue P, Meloni BP, O'Brien PA, Thompson RCA. Molecular characterization of Cryptosporidium isolates from humans and other animals using random amplified polymorphic DNA analysis. Am J Trop Med Hyg 1995;52:559-64. 12. Shianna KV, Rytter R, Spanier JG. Randomly amplified polymorphic DNA PCR analysis of bovine Cryptosporidium parvum strains isolated from the watershed of the Red River of the North. Appl Environ Microbiol 1998;64:2262-5. 13. Bonnin A, Fourmaux MN, Dubremetz JF, Nelson RG, Gobet P, Harly G, et al. Genotyping human and bovine isolates of Cryptosporidium parvum by polymerase chain reaction-restriction fragment length polymorphism analysis of a repetitive DNA sequence. FEMS Microbiol Lett 1996;137:207-11. 14. Widmer G, Tzipori S, Fichtenbaum CJ, Griffiths JK. Genotypic and phenotypic characterization of Cryptosporidium parvum isolates from people with AIDS. J Infect Dis 1998;178:834-40. 15. Spano F, Putignani L, McLauchlin J, Casemore DP, Crisanti A. PCR-RFLP analysis of the Cryptosporidium oocyst wall protein (COWP) gene discriminates between C. wrairi and C. parvum, and between C. parvum isolates of human and animal origin. FEMS Microbiol Lett 1997;150:209-17. 16. Morgan UM, Constantine CC, Forbes DA, Thompson RCA. Differentiation between human and animal isolates of Cryptosporidium parvum using rDNA sequencing and direct PCR analysis. J Parasitol 1997;83:825-30. 17. Peng MM, Xiao L, Freeman AR, Arrowood MJ, Escalante AA, Weltman AC, et al. Genetic polymorphism among Cryptosporidium parvum isolates: evidence of two distinct human transmission cycles. Emerg Infect Dis 1997;3:567-73. 18. Chrisp CE, LeGendre M. Similarities and differences between DNA of Cryptosporidium parvum and C. wrairi detected by the polymerase chain reaction. Folia Parasitol 1994;41:97-100. 19. Carraway M, Tzipori S, Widmer G. Identification of genetic heterogeneity in the Cryptosporidium parvum ribosomal repeat. Appl Environ Microbiol 1996;62:712-6. 20. Kilani RT, Wenman WM. Geographical variation in 18S rRNA gene sequence of Cryptosporidium parvum. Int J Parasitol 1994;24:303-6. 21. Widmer G. Genetic heterogeneity and PCR detection of Cryptosporidium parvum. Adv Parasitol 1998;40:223-39. 22. Cai J, Collins MD, McDonald V, Thompson DE. PCR cloning and nucleotide sequence determination of the 18S rRNA genes and internal transcribed spacer 1 of the protozoan parasites Cryptosporidium parvum and Cryptosporidium muris. Biochim Biophys Acta 1992;1131:317-20. 23. Le Blancq SM, Khramtsov NV, Zamani F, Upton SJ, Wu TW. Ribosomal RNA gene organization in Cryptosporidium parvum. Mol Biochem Parasitol 1997;90:463-78. 24. Johnson DW, Pieniazek NJ, Rose JB. DNA probe hybridization and PCR detection of Cryptosporidium parvum compared to immunofluorescence assay. Water Science and Technology 1993;27:77-84. 25. Johnson DW, Pieniazek NJ, Griffin DW, Misener L, Rose JB. Development of a PCR protocol for sensitive detection of Cryptosporidium oocysts in water samples. Appl Environ Microbiol 1995;61:3849-55. 26. Tzipori S, Griffiths JK. Natural history and biology of Cryptosporidium parvum. Adv Parasitol 1998;40:5-36. 27. Morgan UM, Sargent KD, Deplazes P, Forbes DA, Spano F, Hertzberg H, et al. Molecular characterization of Cryptosporidium from various hosts. Parasitology 1998;117:31-7. 28. Sargent KD, Morgan UM, Elliot A, Thompson RCA. Morphological and genetic characterization of Cryptosporidium oocysts from domestic cats. Vet Parasitol 1998;77:221-7. 29. Navin TR, Weber R, Vugia DJ, Rimland D, Roberts JM, Addiss DG, et al. Declining CD4+ T-lymphocyte counts are associated with increased risk of enteric parasitosis and chronic diarrhea: results of a 3-year longitudinal study. J Acquir Immune Defic Syndr Hum Retrovirol 1999;20:154-9. 30. Melvin DM, Brooke MM. Laboratory procedures for the diagnosis of intestinal parasites. 3rd ed. Atlanta: Centers for Disease Control; 1982. Publication no. (CDC) 85-8282. 31. Garcia LS, Shum AC, Bruckner DA. Evaluation of a new monoclonal antibody combination reagent for the direct fluorescent detection of Giardia cysts and Cryptosporidium oocysts in human fecal specimens. J Clin Microbiol 1992;30:3255-7. 32. da Silva AJ, Bornay-Llinares FJ, Moura INS, Slemenda SB, Tuttle JL, Pieniazek NJ. Fast and reliable extraction of protozoan parasite DNA from fecal specimens. Molecular Diagnosis. In press 1999. 33. Schuler GD, Altschul SF, Lipman DJ. A workbench for multiple alignment construction and analysis. Proteins 1991;9:180-90. 34. Felsenstein J. PHYLIP—phylogeny inference package. Cladistics 1989;5:164-6. 35. Roth A, Fischer M, Hamid ME, Michalke S, Ludwig W, Mauch H. Differentiation of phylogenetically related slowly growing mycobacteria based on 16S-23S rRNA gene internal transcribed spacer sequences. J Clin Microbiol 1998;36:139-47. 36. Bornay-Llinares FJ, da Silva AJ, Moura IN, Myjak P, Pietkiewicz H, Kruminis-Lozowska W, et al. Identification of Cryptosporidium felis in a cow by morphologic and molecular methods. Appl Environ Microbiol -------------------------------------------------------------------------- Dispatches Fatal Case Due to Methicillin-Resistant Staphylococcus aureus Small Colony Variants in an AIDS Patient Harald Seifert,* Christoph von Eiff,† and Gerd Fätkenheuer† *University of Cologne, Cologne, Germany; and †Westfälische Wilhelms-Universität, Münster, Germany --------------------------------------------------------------------------- We describe the first known case of a fatal infection with small colony variants of methicillin-resistant Staphylococcus aureus in a patient with AIDS. Recovered from three blood cultures as well as from a deep hip abscess, these variants may have resulted from long-term antimicrobial therapy with trimethoprim/sulfamethoxazole for prophylaxis of Pneumocystis carinii pneumonia. Staphylococcus aureus causes acute and often fatal infections. Small colony variants (SCVs), which are subpopulations of S. aureus, are implicated in persistent and recurrent infections (in particular osteomyelitis, septic arthritis, respiratory tract infections in patients with cystic fibrosis, and deep-seated abscesses) (1-4). These phenotypic variants produce small, slow-growing, nonpigmented, nonhemolytic colonies on routine culture media, making correct identification difficult for clinical laboratories. Biochemical characterization of these variants suggests that they are deficient in electron transport activity (5). We report a fatal case of a persistent deep-seated hip abscess due to methicillin-resistant S. aureus SCVs that led to osteomyelitis and bloodstream infection in a patient with AIDS. Case Report A 36-year-old man with AIDS came to the Cologne University Hospital, Cologne, Germany, in June 1997 with fever and progressive pain (of 6 weeks duration) in his right hip. HIV infection had been diagnosed in 1986. In 1994, his CD4 cell count was 250/µL, and oral zidovudine therapy was started. His medical history included Pneumocystis carinii pneumonia, pulmonary tuberculosis, and recurrent oral thrush; his medication included zidovudine, lamivudine, fluconazole, and trimethoprim/sulfamethoxazole. In September 1996, he was in a traffic accident and had severe cerebral trauma resulting in spastic hemiparesis with occasional seizures. After an intramuscular injection 2 months before admission, pus was surgically drained to treat recurrent abscesses of his right hip. Specimens for culture were not obtained. Physical examination found limited mobility of his right thigh and a tender, nondraining scar at the site of surgical drainage. Neither warmth nor swelling was observed over his right hip. Vital signs were temperature, 38.2°C; respiration rate, 28; and heart rate, 108. He was awake and alert and had spastic paresis in his right arm. Laboratory studies performed on admission showed hemoglobin, 10.8 g/dL; leukocyte count, 3,000/µL with a normal differential; CD4 cell count, 20/µL; platelet count, 131,000/µL; C-reactive protein, 184 mg/L; and alkaline phosphatase, 1490 U/L. Radiographs of the chest and a plain film of the pelvis were normal. A triple-phase bone scan showed an area of minor tracer accumulation in the acetabulum region of the right hip. Blood cultures were drawn, but antimicrobial therapy was withheld until culture results became available. On hospital day 2, one of two blood cultures drawn on admission yielded nonhemolytic staphylococci that were clumping factornegative. The organisms were initially misidentified as coagulase-negative staphylococci and were considered contaminants. Empiric antistaphylococcal therapy with clindamycin (600 mg q8hr) was instituted. On hospital day 4, two sets of blood cultures obtained on hospital day 2 yielded phenotypically identical organisms, which on the basis of a positive tube coagulase test were identified as oxacillin-resistant S. aureus. The colony morphology was suggestive of an SCV of S. aureus. The patient was started on parenteral vancomycin treatment (1 g q12hr). However, his condition deteriorated rapidly, and he died of refractory septic shock 6 days after admission. Autopsy showed a large (12 x 10 x 8 cm), deep-seated abscess of the right hip and osteomyelitis of the ischial tuberosity. Both SCVs and typical large colony forms of S. aureus were cultured from postmortem specimens of the abscess and the bone. Findings S. aureus SCVs were recovered from one of two blood culture sets obtained on admission and from two of four blood culture sets obtained on hospital day 2. Growth was not detected until the blood culture bottles had been incubated 24 hours. S. aureus with a normal phenotype was recovered from nose and throat specimens but not from blood cultures, whereas both SCVs and typical S. aureus phenotypes were isolated from the deep hip abscess (Figure 1) before death, as well as from a postmortem specimen. All isolates were clumping factor–negative but showed a delayed positive reaction in the tube-coagulase test at 24 hours. The results of the ID 32 staph test did not unambiguously identify SCVs as S. aureus because the tests for urease and trehalose were negative. Both the nuc gene and the coa gene were identified by polymerase chain reaction (PCR) amplification. Methicillin resistance was confirmed for both small and large colony forms by PCR amplification of the mecA gene. When cultured without supplementation, all SCVs were nonpigmented and nonhemolytic. Supplementation with hemin, thymidine, or menadione identified two SCVs showing thymidine auxotrophy and a combined thymidine and menadione auxotrophy, respectively. All SCVs were stable on repeated subculturing. Epidemiologic typing by PCR analysis of inter-IS256 spacer length polymorphisms (Figure 2) and pulsed-field gel electrophoresis of genomic DNA (data not shown) showed identical banding patterns for both SCVs and large colony forms, which indicates that the phenotypically different S. aureus isolates represented a single strain. Antimicrobial susceptibility testing was performed by microbroth dilution, according to the National Committee for Clinical Laboratory Standards guidelines. Susceptibility to trimethoprim/sulfamethoxazole was tested with Etest (AB Biodisk, Solna, Sweden). In contrast to current standards, the MICs for SCVs were determined after 48 hours of incubation at 35°C. Susceptibility testing showed that all S. aureus isolates were resistant to penicillin (MIC, >8 µg/mL), ampicillin (MIC, >32 µg/mL), oxacillin (MIC, >8 µg/mL), erythromycin (MIC, >32 µg/mL), clindamycin (MIC, >32 µg/mL), ciprofloxacin (MIC, >8 µg/mL), gentamicin (MIC, >500 µg/mL), and trimethoprim/sulfamethoxazole (MIC, >32 µg/mL) and susceptible to vancomycin (MICs, 1-2 µg/mL), teicoplanin (MICs, 0.5-1 µg/mL), and quinupristin/dalfopristin (MICs, 0.5-1 µg/mL). No differences in MICs were observed between S. aureus SCVs and S. aureus isolates with normal phenotype. [Fig] Figure 1. Staphylococcus aureus small colony variants(A) and S. aureus with a normal phenotype(B) cultured on sheep blood agar after 24 hours of incubation at 35°C. Staphylococci were identified by conventional methods(6) and with the ID 32 Staph system (bioMérieux, Marcy-L'Etoile, France) following the instructions of the manufacturer. The tube-coagulase test was read after 24 hours. S. aureus isolates were characterized as small colony variants as described before(7-8). Auxotrophic requirements were evaluated with 10-µg hemin disks, 1.5-µg menadione disks, and 1.5-µg thymidine disks on Mueller-Hinton agar and on chemically defined medium (CDM) agar as well as on CDM agar supplemented with 1 µg/mL hemin, 100 µg/mL thymidine, and 1 µg/mL menadione, menadione, [Fig] Figure 2. Fingerprint patterns obtained for Staphylococcus aureus small colony variants (lanes 3-5, bloodculture isolates; lanes 6 and 7, isolates from hip abscess; lane 8, postmortem specimen) and S. aureus isolates with a normal phenotype (lanes 10 and 11, isolates from nose and throat; lanes 12 and 13,isolates from hip abscess and postmortem specimen) after polymerase chain reaction (PCR) analysis of inter-IS256 spacer length showing identical strains. Lane 1, 100-bp ladder; lanes 2, 9, and 16, methicillin-resistant S. aureus (MRSA) reference strain; lanes 14 and 15, epidemiologically unrelated MRSA strains. Strain relatedness of all isolates with different colony morphologies and from different sources was analyzed by PCR analysis of inter-IS256 spacer length polymorphisms (9) and pulsed-field gel electrophoresis after SmaI restriction (8). Minor modifications included the use of brain heart infusion broth instead of trypticase soy broth to obtain sufficient growth of S. aureus small colony variants. To our knowledge, this case represents the first of a serious S. aureus infection in an AIDS patient in which all blood cultures yielded SCVs. The SCVs' unusual morphologic appearance and slow growth delayed the correct identification of these organisms as S. aureus. The empiric antimicrobial regimen in our patient did not include a glycopeptide, because of the low rate of methicillin resistance in community-acquired S. aureus infection in Germany. Appropriate antistaphylococcal therapy was, therefore, not started until hospital day 4. Delayed antimicrobial therapy on day 4 rather than on day 2 may have contributed to the patient's death. Proctor and colleagues recently reported five cases in which SCVs of S. aureus were implicated in persistent and relapsing infections. They identified only a single case reported in the previous 17 years and ascribed this to insufficient ability of laboratories to identify these organisms (8). In most cases, patients had received antibiotics. Aminoglycoside treatment may have selected for S. aureus SCVs (10), and in cases of osteomyelitis or deep-seated abscesses, persistence of these variants in the intracellular milieu may have permitted evasion of host defenses and allowed for the development of resistance to antimicrobial therapy (7,11). Von Eiff and colleagues recently reported four cases of chronic osteomyelitis due to SCVs of S. aureus in patients who had received gentamicin beads as an adjunct to surgical therapy for osteomyelitis (2). Kahl et al. described persistent infection with S. aureus SCVs in patients with cystic fibrosis (4). All these patients had received long-term trimethoprim/sulfamethoxazole prophylaxis. It may be tempting to speculate that administration of trimethoprim/sulfamethoxazole for prophylaxis against P. carinii pneumonia may have selected for SCVs within the patient's large hip abscess. Further prospective studies are needed to assess the role of S. aureus SCVs in HIV-infected patients on long-term antimicrobial therapy. --------------------------------------------------------------------------- Dr. Seifert is assistant professor at the Institute of Medical Microbiology and Hygiene, University of Cologne, Germany. His research interests include the molecular epidemiology of nosocomial pathogens, in particular Acinetobacter species, catheter-related infections, and antimicrobial resistance. Address for correspondence: Harald Seifert, Institute of Medical Microbiology and Hygiene, University of Cologne, Goldenfelsstraße 19-21, 50935 Cologne, Germany; fax: 49-221-478-3081; e-mail: harald.seifert@uni-koeln.de. References 1. Proctor RA, Balwit JM, Vesga O. Variant subpopulations of Staphylococcus aureus as cause of persistent and recurrent infections. Infectious Agents and Disease 1994;3:302-12. 2. von Eiff C, Bettin D, Proctor RA, Rolauffs B, Lindner N, Winkelmann W, et al. Recovery of small colony variants of Staphylococcus aureus following gentamicin bead placement for osteomyelitis. Clin Infect Dis 1997;25:1250-1. 3. Spearman P, Lakey D, Jotte S, Chernowitz A, Claycomb S, Stratton C. Sternoclavicular joint septic arthritis with small colony variant Staphylococcus aureus. Diagn Microbiol Infect Dis 1996;26:13-5. 4. Kahl B, Herrmann M, Everding AS, Koch HG, Becker K, Harms E, et al. Persistent infection with small colony variant strains of Staphylococcus aureus in patients with cystic fibrosis. J Infect Dis 1998;177:1023-9. 5. von Eiff C, Heilmann C, Proctor RA, Woltz C, Peters G, Götz F. A site-directed Staphylococcus aureus hemB mutant is a small colony variant which persists intracellularly. J Bacteriol 1997;179:4706-12. 6. Kloos WE, Bannerman TL. Staphylococcus and micrococcus. In: Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, editors. Manual of clinical microbiology. 6th ed. Washington: American Society for Microbiology; 1995. p. 282-98. 7. Balwit JM, van Langevelde P, Vann JM, Proctor RA. Gentamicin-resistant menadione and hemin auxotrophic Staphylococcus aureus persist within cultured endothelial cells. J Infect Dis 1994;170:1033-7. 8. Proctor RA, van Langevelde P, Kristjansson M, Maslow JN, Arbeit RD. Persistent and relapsing infections associated with small colony variants of Staphylococcus aureus. Clin Infect Dis 1995;20:95-102. 9. Deplano A, Vaneechoutte M, Verschraegen G, Struelens MJ. Typing of Staphylococcus aureus and Staphylococcus epidermidis by PCR analysis of inter-IS256 spacer length polymorphisms. J Clin Microbiol 1997;35:2580-7. 10. Pelletier LL Jr, Richardson M, Feist M. Virulent gentamicin-induced small colony variants of Staphylococcus aureus. J Lab Clin Med 1979;94:324-34. 11. Proctor RA, Kahl B, von Eiff C, Vaudaux PE, Lew DP, Peters G. Staphylococcal small colony variants have novel mechanisms for antibiotic resistance. Clin Infect Dis 1998;27 Suppl 1:S68-74. -------------------------------------------------------------------------- Dispatches Application of Data Mining to Intensive Care Unit Microbiologic Data(ft 1) Stephen A. Moser, Warren T. Jones, and Stephen E. Brossette The University of Alabama at Birmingham, Birmingham, Alabama, USA --------------------------------------------------------------------------- We describe refinements to and new experimental applications of the Data Mining Surveillance System (DMSS), which uses a large electronic health-care database for monitoring emerging infections and antimicrobial resistance. For example, information from DMSS can indicate potentially important shifts in infection and antimicrobial resistance patterns in the intensive care units of a single health-care facility. We have defined a new exploratory data mining process for automatically identifying new, unexpected, and potentially interesting patterns in hospital infection control and public health surveillance data. This process, and the system based on it, Data Mining Surveillance System (DMSS), use association rules to represent outcomes and association rule confidences to monitor changes in the incidence of those outcomes over time. Through experiments with infection control data from the University of Alabama at Birmingham Hospital, we have demonstrated that DMSS can identify potentially interesting and previously unknown patterns. Future work on prospective clinical studies to determine the usefulness of DMSS in hospital infection control is needed, as is improved event presentation for the user and strategies for handling larger datasets. The statistical strategies developed for automatically detecting temporal patterns in surveillance data require that analysts explicitly define outcomes of interest before surveillance begins. The Data Mining Surveillance System (DMSS), on the other hand, is not constrained to monitoring changes in user-defined outcomes. In DMSS, complex outcomes are represented by association rules, and outcome incidence is captured monthly. An early version of DMSS, along with association rules and early experiments with a single organism, has been described (1). We briefly describe a newer version of DMSS and experimental results obtained by using it to analyze 1 year's data from intensive care units (ICUs) at the University of Alabama at Birmingham Hospital. DMMS uses the following definitions. An itemset is a subset of the set of all items. The support of an itemset x, sup (x), is the number of records that contain x. If sup (x) >/= FSST, where FSST is the frequent set support threshold (FSST), then x is a frequent set. An association rule, A ==> B, where A and B are frequent sets and the insection of A and B = Ø, is a is a statement about how often the items of B are found with the items of A. the incidence proportion of A ==> B, denoted ip(A ==> B), is equal to sup (union of A and B)/sup (A). The precondition support of association rule A ==> B is sup(A). The incidence proportion of an association rule A ==> B in data partition p(sub i)describes the incidence of the outcome, B, in the group, A, during time ti. A series of incidence proportions for A ==> B from partitions p(sub1), p(sub 2), ...., p(sub n)describes the incidence of the outcome B in group A from t(sub 1) through t(sub n). Therefore, by analyzing the series of incidence proportions of an association rule A==> B, it should be possible to detect important shifts or trends in the incidence of B in A over time. In this way, surveillance of B in A is possible. Bacterial susceptibility and related demographic data of patients in the University of Alabama at Birmingham Hospital ICUs (medical, surgical [SICU], cardiac, neurologic [NICU]) during 1997 were extracted from the PathNet laboratory information system. Each record describes a single isolate and contains the following data elements: date of admission, date of sample collection, date of results reported, source of isolate (e.g., sputum, blood), organism isolated, organism Gram stain and morphologic features, patient's location in the hospital, and resistant (R), intermediate (I), or susceptible (S) test results to relevant antibiotics, according to the National Committee for Clinical Laboratory Standards MIC breakpoints (2). Duplicate records were removed so that for each patient, no more than one isolate per organism per month was included. In each remaining record, certain antimicrobial drug items were removed (only drugs to which the organism is historically susceptible at least 50% of the time remained). Additionally, items of the form S~Antimicrobial were removed so that only I~Antimicrobial and R~Antimicrobial items remained. Finally, data were divided into 1-month partitions (p(sub 1)....p(sub n)) before analysis. For each partition p(sub i), all frequent sets with support of at least 3 (FSST >2) and association rules with precondition support greater than 5 were generated. Both the frequent set discovery and association rule- generating algorithms are beyond the scope of this review (3). Each generated association rule must pass a set of rule templates that describe families of interesting and uninteresting rules. Each template is a construct of the form be(sub 1) ==> be(sub 2), where be(sub 1) and be(sub 2) are Boolean expressions over items and attributes. Association rule A ==> B satisfies rule template be(sub 1) ==> be(sub 2) if A satisfies be1 and B satisfies be(sub 2). Two types of association rule templates are used: include templates and exclude templates. An association rule A ==> B passes a set of rule templates if A ==> B satisfies at least one include template in the set and does not satisfy any exclude template in the set. Rule templates are handcrafted by domain experts to eliminate inherently uninteresting or nonsense rules. This is accomplished through iterative experiments with representative data by initially using few templates and then creating and modifying templates on the basis of pattern review. History is a database that holds association rules and their incidence proportions for different data partitions. In DMSS, the user specifies a set of rule templates that contains any number of inclusive and restrictive templates (Table 1). Only association rules that pass the rule templates are included in the history. To establish a baseline for an association rule, the incidence proportions of the rule for the three previous partitions are obtained and stored in the history. Once stored in the history, a rule is updated for each new partition regardless of whether or not it is generated in the partition. Therefore, for every association rule, the history contains an up-to-date time-series of incidence proportions. Table 1. Templates used to filter association rules -------------------------------------------------------------------------- Template type Left (be(sub 1)) Right (be(sub 2)) Explanation -------------------------------------------------------------------------- Exclude (R~Antibiotic) (Anything) Want antibiotic sensitivity info on the right only. Exclude (Anything) (Source) Source of infection is not an outcome. Therefore, exclude all rules with a source on the right. Exclude (NS OR Org (NS OR Org NS, Org, and GrMp are GrMP) OR GrMP more informative if kept together in either a group or an outcome. Exclude (Loc) (Org OR GrMp) If the left contains AND location, then exclude rules that (R~Antibiotic) have Org and R~Antibiotic or GrMp and R~Antibiotic. Include (Org OR Loc) (R~Antibiotic OR Include rules whose GrMp OR Org) groups are Org- or AND Not (Loc) Loc-specific and whose outcomes are Antibiotic- or GrMp-specific. -------------------------------------------------------------------------- be(sub 1) and be(sub 2), Boolean expressions; R, resistant; NS, nosocomial; OR, "or"; Org, organism; GrMp, Gram stain and morphology; Loc, Location. Table 2. A sample event generated by the Data Mining Surveillance System ----------------------------------------------------------------------------- Association P P P P P P rule (subc-5)(subc-4)(subc-3)(subc-2)(subc-1)(subc) (sup a) ----------------------------------------------------------------------------- (nosocomial ==> {Acinetobacter 0/11 0/10 0/9 0/13 2/9 3/9 SICU(sup b), baumannii} tracheal aspirate ----------------------------------------------------------------------------- w(subp) w(subc) (sup c) ----------------------------------------------------------------------------- (sup a)P(subc), current pair. (sup b)SICU, surgical intensive care unit. (sup c)w(subp), past window; w(subc), current window. By analyzing information stored in the history, DMSS generates alerts that describe an extreme change in the incidence of an outcome B in a group A over time. For example, Table 2 describes the incidence of Acinetobacter baumannii in a nosocomial tracheal aspirate and in SICU isolates over the past six partitions. Clearly, a shift in incidence occurs between the first 4 months and the most recent 2 months of the series. If we call months 1, 2, 3, and 4 the past window, wp, and months 5 and 6 the current window, w(sub c), we can ask if there is an extreme change in the incidence between w(sub p) and w(sub c). We compute the cumulative incidence proportion for w(sub p) (0/43) and for w (sub c)(5/18) and compare the two by a statistical test of two proportions. To generate an alert for an association rule r, DMSS first constructs a current window (w(sub c)) and a past window (w(sub p)) on the series of incidence proportions of r (w(sub c)[r,0], w (sub p)[r,0] from the algorithm in the Figure). Second, it computes the cumulative incidence proportion for each window. Third, it compares the two cumulative incidence proportions by a test of two proportions. Finally, if the difference between the proportions is statistically extreme (p R~Oxacillin 0/10 0/8 7/14 Increase in the aureus (sup a,b) incidence of Source R~Clindamycin oxacillin (ORSA), TRACHASP(sup c) R~Erythromycin clindamycin and erythromycin resistance in all isolated from tracheal aspirates. NSNoso(sup d) ==> R~Ceftazidime 3/88 11/70 Increase in incidence of ceftazidime resistance in all nosocomial isolates. NP_GNR(sup e) ==> R~Piperacillin 0/17 6/14 Increase in the LocSICU incidence of piperacillin resistance in non-pseudomonas gram-negative bacilli isolated from NSNoso. NP_GNR ==> R~Piperacillin 1/12 0/14 4/11 4/8 Increase in the LocSICU (sup f) incidence of piperacillin resistance in non-pseudomonas, nosocomial gram- negative bacilli from the SICU. NSNoso ==> S. aureus 26 3/26 2/28 6/27 5/20 3/11 Increase in the LocNICUg incidence of nosocomial S. aureus in nosocomial isolates from the NICU. ------------------------------------------------------------------------------ (sup a)R, resistant. (sup b)Oxacillin, resistance implies resistance to amoxycillin/clavulanic acid, cephalothin, and cefazolin. (sup c)SourceTRACHASP, tracheal aspirates. (sup d)NSNoso, nosocomial (3 days from admission). (sup e)NP_GNR, non-pseudomonas gram-negative rod. (sup f)LocSICU, location, surgical intensive care unit (SICU). (sup g)LocNICU, location, neonatal intensive care unit (NICU). We believe that this approach to surveillance will allow hospital infection control programs to focus their limited resources on issues of probable significance. We also believe that this approach is a step toward the public health surveillance system described by Dean, Fagan, and Panter-Conner (4). --------------------------------------------------------------------------- This work was supported in part by cooperative agreement U47-CCU411451 with the Centers for Disease Control and Prevention (SAM) and a predoctoral research fellowship LM-00057 from the National Library of Medicine (SEB). Dr. Moser is associate professor, Department of Pathology, University of Alabama at Birmingham, and serves as director of Laboratory Information Services, associate director of Clinical Microbiology for University Hospital, and director of the Pathology Informatics Section. His research interests are applied research in diagnostic microbiology and the application of software as an aid to the intelligent analysis of medical information, especially that generated in laboratory medicine. Address for correspondence: Stephen A. Moser, University of Alabama at Birmingham, Department of Pathology, P246, 619 19th St., South Birmingham, AL 35233-7331, USA; fax: 205-975-4468; e-mail: moser@uab.edu. (footnote 1)Presented in part at the International Conference on Emerging Infectious Diseases, March 8-11, 1998, Atlanta, Georgia. References 1. Brossette SE, Sprague AP, Hardin JM, Waites KB, Jones WT, Moser SA. Association rules and data mining in hospital infection control and public health surveillance. J Am Med Inform Assoc 1998;5:373-81. 2. National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. 4th ed. Approved standard. NCCLS document M7-A4. Wayne (PA): The Committee; 1997. 3. Brossette SE. Data mining and epidemiologic surveillance [dissertation]. Birmingham (AL): University of Alabama at Birmingham; 1998. 4. Dean AG, Fagan RF, Panter-Conner BJ. Computerizing public health surveillance systems. In: Teutsch SM, Churchill RE, editors. Principles and practice of public health surveillance. New York: Oxford University Press; 1994. p. 200-17. -------------------------------------------------------------------------- Dispatches Sentinel Surveillance for Enterovirus 71, Taiwan, 1998 Trong-Neng Wu,* Su-Fen Tsai,* Shu-Fang Li,* Tsuey-Fong Lee,* Tzu-Mei Huang,* Mei-Li Wang,* Kwo-Hsiung Hsu,† and Chen-Yang Shen‡ *Disease Surveillance and Quarantine Service, Ministry of Health, Taiwan, Republic of China; †Bureau of Communicable Disease Control, Ministry of Health, Taiwan, Republic of China; and ‡Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan, Republic of China --------------------------------------------------------------------------- Outbreaks of enterovirus 71 have been reported around the world since 1969. The most recent outbreak occurred in Taiwan during April-July 1998. This hand, foot, and mouth disease epidemic was detected by a sentinel surveillance system in April at the beginning of the outbreak, and the public was alerted. Enterovirus type 71 (EV71), one of the etiologic agents of epidemic hand, foot, and mouth disease (HFMD), has been associated with febrile rash illness, aseptic meningitis, encephalitis, and a syndrome of acute flaccid paralysis similar to that caused by poliovirus (1,2). EV71 was identified in 1969 in the United States, when it was isolated from the feces of an infant with encephalitis in California. By 1998, many EV71 outbreaks had been reported around the world. In addition to the outbreak in California in which one death was reported, four other outbreaks resulted in many fatal cases involving clinical deterioration and death in young children (Bulgaria, May-September 1975; Hungary, 1978; Malaysia, April-June 1997; Taiwan, April-August 1998) (3-5). To the best of our knowledge, the outbreak in Taiwan marked the first time that an EV71 outbreak was detected by a surveillance system, which alerted the public about the epidemic of HFMD. We describe how the EV71 outbreak was reported by a sentinel surveillance system established in July 1989 by Disease Surveillance and Quarantine Service (originally National Quarantine Services), Ministry of Health, Taiwan. In this sentinel surveillance system, public health officers contact local and regional physicians weekly to actively collect disease information. On the basis of information collected, disease incidence trends are predicted, and, if necessary, the public is warned during the earliest stage of disease outbreaks. Mumps, varicella, diarrhea, and upper-tract respiratory infection are among the diseases subject to routine surveillance. For convenience, we have established two channels of data collection: telephone interviews and report cards mailed by physicians. Approximately 850 physicians (fewer than one tenth of the physicians) from every county in Taiwan participate in the system: 258 in the northern region, 211 in the central region, 296 in the southern region, and 85 in the eastern region (Table); most of these are pediatricians, general practitioners, and family physicians. Only a few are ear, nose, and throat specialists. Table. Physician distribution, Sentinel When an epidemic of fatal Surveillance System, Taiwan myocarditis was reported in Sarawak, Malaysia, in 1997, we -------------------------------------------- started to collect information No. cards regarding HFMD and vesicular Region of mailed by Total no. pharyngitis (herpangina) Taiwan Interviews physicians physicians through the sentinel -------------------------------------------- surveillance system. Beginning in March 1998, some physicians Northern 109 149 258 reported a notable increase of Central 27 184 211 cases of HFMD, vesicular Southern 175 121 296 pharyngitis (herpangina), and vesicular stomatitis exanthem. Eastern 13 72 85 Furthermore, a dramatic Total 324 526 850 upsurge of HFMD was seen in children in outpatient -------------------------------------------- settings at the end of April 1998 (Figure 1). Surveillance data provided the basis for immediate public health action. When the number of weekly reported cases increased twofold at the end of April, the public was informed about the epidemic of HFMD and herpangina and the threat of enterovirus infection (May 12). Measures for preventing the spread of infection (e.g., practicing good hygiene at all times, confining infected children at home, avoiding contact with infected children) were advised. However, the number of reported cases still dramatically increased to approximately 3,000 in the following week. Although some errors in reporting might have occurred, we are confident that the reporting quality was adequate. The sentinel physicians, who have voluntarily cooperated with us for 10 years, used a clinical case definition of HFMD/herpangina we created when the surveillance started. Monitoring incidence of the [fig] [fig] fatal and most severe Figure 1. Total cases of Figure 2. Number of cases of hand, foot, and mouth disease hospitalizations and HFMD and herpangina reported from severe cases of hand, appeared sentinel physicians in foot, and mouth disease critical; Taiwan, March 19 to August and herpangina in Taiwan, therefore, 29, 1998. June-August, 1998. another report system, designed for monitoring severe and fatal cases, was established (May 29) to enroll all well-defined severe and fatal cases in 597 hospitals and medical centers. This new system established a network of various public health agencies, general and regional hospitals, and medical centers. The difference between the two systems was that, while the sentinel surveillance system was physician-based, the new system was hospital-based. Severe and fatal cases were defined as HFMD with complications, including aseptic meningitis, encephalitis, myocarditis, acute flaccid paralysis, rapidly deteriorating clinical course, and death. Both surveillance systems worked simultaneously from June 1998 onward. We found that the trend peaked and declined earlier in the sentinel system than in the hospital-based surveillance system (Figure 1, 2). A case-control study was implemented, and enterovirus isolation data were reviewed. Several enterovirus isolations, from patients with severe and fatal cases, were in stool specimens, throat secretions, cerebrospinal fluid, blood, and central nervous system tissue, including EV71, Coxsackie, and ECHO. Additional studies comparing rates of EV71 isolation in different years are in progress and will be reported separately. Most isolated viruses were EV71 from all specimens in this epidemic. The epidemiologic, clinical, and virologic evidence suggests an association between EV71 infection and this epidemic of HFMD. However, the causes of the severe cases and deaths in Taiwan are yet to be defined (5). A physician-based sentinel surveillance system can play an important role in preventing emerging infectious diseases. Even though the data collected may be rough, a sentinel surveillance system can provide necessary information for monitoring communicable diseases, guiding further investigation, and evaluating control measures, as well as early warning for epidemics and rationale for public health intervention. Early detection of communicable diseases and immediate public health intervention can curtail the number of illnesses and deaths and reduce negative effects on international travel and trade (6). Dr. Wu is vice superintendent of Municipal Hsiao-Kang Hospital, Kaohsiung Medical College, Taiwan, Republic of China. He is also a professor at both Kaohsiung Medical College, Kaosiung, Taiwan, and China Medical College, Taichung, Taiwan. Address for correspondence: Su-Fen Tsai, Division of Disease Surveillance, Disease Surveillance and Quarantine Service, Ministry of Health, F6, No.6, Lin Shen S. Rd., Taipei, Taiwan, R.O.C.; fax: 886-2-23945312; e-mail: sftsai@net.dsqs.gov.tw. References 1. Melnick JL. Enterovirus type 71 infection: a varied clinical pattern sometimes mimicking paralytic poliomyelitis. Rev Infect Dis 1984;6 Suppl:S387-90. 2. Alexander JP Jr, Baden L, Pallansch MA, Anderson LJ. Enterovirus 71 infection and neurologic disease—United States, 1977-1991. J Infect Dis 1994;169:905-8. 3. Shindarov LM, Chumakov MP, Voroshilova MK, Bojinov S, Vasilenko SM, Iordanov I, et al. Epidemiological, clinical, and pathomorphological characteristics of epidemic poliomyelitis-like disease caused by enterovirus 71. Journal of Hygiene, Epidemiology, Microbiology, and Immunology 1979;23:284-95. 4. Nagy G, Takatsy S, Kukan E, Mihaly I, Domok I. Virological diagnosis of enterovirus type 71 infections: experiences gained during an epidemic of acute CNS diseases in Hungary in 1978. Arch Virol 1982;71:217-27. 5. Centers for Disease Control and Prevenetion. Deaths among children during an outbreak of hand, foot, and mouth disease—Taiwan, Republic of China, April-July 1998. MMWR Morb Mortal Wkly Rep 1998;47:629-32. 6. Heymann DL, Rodier GR. Global surveillance of communicable diseases. Emerg Infect Dis 1998;4:362-5. -------------------------------------------------------------------------- Dispatches Chlorine Inactivation of Escherichia coli O157:H7 Eugene W. Rice, Robert M. Clark, and Clifford H. Johnson U.S. Environmental Protection Agency, Cincinnati, Ohio, USA --------------------------------------------------------------------------- We analyzed isolates of Escherichia coli O157:H7 (which has recently caused waterborne outbreaks) and wild-type E. coli to determine their sensitivity to chlorination. Both pathogenic and nonpathogenic strains were significantly reduced within 1 minute of exposure to free chlorine. Results indicate that chlorine levels typically maintained in water systems are sufficient to inactivate these organisms. Escherichia coli O157:H7 is becoming increasingly recognized as a waterborne pathogen. Two recent outbreaks during summer 1998, one involving a drinking water supply in Wyoming (1) and another involving recreational water exposure at a water park in Georgia (2), have underscored the role of water in transmission. Contaminated drinking water (3,4) and recreational water have been associated with outbreaks of hemorrhagic colitis caused by E. coli O157:H7 (5-7). Chlorination of water is one of the primary public health measures used to ensure that both potable water and water used in recreational settings are free of microbial pathogens. Our study was undertaken to determine the chlorine resistance of E. coli O157:H7 and compare this resistance with that of wild-type E. coli. Seven strains of E. coli O157:H7, isolated from cattle from geographically distinct areas (Florida, Idaho, Illinois, Missouri, Texas, Washington, and Wisconsin), were obtained from the U.S. Department of Agriculture (D. Miller, Ames, IA). The isolates exhibited the characteristic phenotypic traits: sorbitol-negative, ß-glucuronidase–negative, lactose-positive, indole-positive, and positive for glutamate decarboxylase (8). All enterohemorrhagic isolates were active toxin producers, as determined by in vitro enzyme immunoassay (Meridian Diagnostics, Inc., Cincinnati, OH). These cattle isolates were chosen as representative strains that might contaminate water supplies after surface run-off from pastures and fields. Four wild-type E. coli isolates from cattle manure from a local dairy farm (Ohio) were characterized by biochemical test kits (bioMerieux Vitek, Hazelwood, MO). All bacterial cultures used in the disinfection experiments were grown for 18 to 20 hours at 35ºC in brain heart infusion broth, concentrated by centrifugation, and washed three times in phosphate buffer (9) before testing. The results of the disinfection experiments, including the rates of inactivation, are shown in the Table. Initial levels for all isolates were 5.52 to 5.79 log(sub 10) CFU/ml. The mean chlorine levels at each exposure time were 1.1 mg/L free chlorine and 1.2 mg/L total chlorine. For both the pathogenic and the wild-type strains, exposure to these levels of chlorine for 1 minute reduced the viable populations by approximately four orders of magnitude. The inactivation rates and corresponding correlation coefficient (r[sup 2]) values are listed in the Table. Little difference was observed in the rates of inactivation for the pathogenic and wild-type organisms. Table. Chlorine inactivation of Escherichia coli O157:H7 and wild-type E. coli(sup a) --------------------------------------------------------------------------- Log(sub 10)CFU/ml -------------------------------- After exposure time of ------------------------ Initial Inactivation Isolate inoculum 30 sec 60 sec 120 sec rate(sec[sup -1])r(sup 2) --------------------------------------------------------------------------- E. coli O157:H7 N009-6-1 5.63 2.60 1.88 0.82 -2.96 0.82 N6001-8-10 5.78 2.52 1.44 0.72 -3.06 0.68 N6021-5-1 5.78 2.54 1.52 0.66 -3.06 0.54 N60049-26-1 5.68 2.35 1.40 0.54 -3.00 0.86 N6059-7-2 5.72 2.42 1.74 0.86 -3.02 0.72 N6104-5-9 5.62 2.40 1.69 0.72 -2.96 0.89 N6114-7-2 5.63 2.52 1.66 0.89 -2.96 0.82 Mean 5.69 2.48 1.62 0.74 -2.93 0.82 E. coli (wild type) A 5.53 2.66 1.80 1.52 -2.51 0.61 B 5.79 2.60 1.48 0.81 -2.68 0.60 C 5.68 2.48 0.92 0.84 -2.61 0.61 D 5.52 2.34 0.95 0.39 -2.50 0.61 Mean 5.63 2.52 1.28 0.89 -2.93 0.71 --------------------------------------------------------------------------- (sup a)In chlorine demand-free chlorinated (CDF) buffer, 5ºC, pH 7.0, 1.1 mg/L free chlorine, 1.2 mg/L total chlorine. Duplicate chlorine inactivation experiments were conducted in CDF buffer at pH 7.0. All experiments were conducted at 5ºC in a recirculating, refrigerated water bath. The chlorinated buffer was prepared by the addition of reagent-grade sodium hypochlorite (Fisher Scientific, Fair Lawn, NJ). Reaction vessels were continuously mixed (250 rpm) by using an overhead stirring apparatus equipped with sterile stainless steel paddles. Chlorine concentrations were determined by the N,N-dimethyl-p-phenylenediamine colorimetric method (9). Samples were removed from the reaction vessels at the desired exposure times, and the chlorine was immediately neutralized by the addition of 0.5 ml of 10% (wt/vol) sodium thiosulphate. Vessels containing CDF buffer without chlorine served as controls for determining unexposed concentrations of the bacteria. Initial levels and the number of survivors after chlorine exposure were determined by the membrane filtration procedure using mT7 agar incubated for 22 to 24 hours at 35ºC. This medium was chosen because of its ability to recover oxidant-stressed organisms (9). Levels of bacteria were determined by duplicate filtrations of appropriate dilutions for each exposure time.The log(sub 10)-transformed data were used to determine the levels of inactivation for each isolate. The means for the inactivation data for the E. coli O157:H7 isolates and for the wild-type E. coli isolates at each exposure time were used to compare the inactivation rates between the pathogenic and the wild-type organisms. The following first order model was used to describe the inactivation rate: y = y(sub 10)10(sup -at), where t = time in seconds, y = CFU/ml at any time t, y(sup 10)= CFU/ml at time zero, and a = the inactivation rate in sec(sup -1). The log transformation of this equation was used to calculate the inactivation rate. A regression analysis using least squares was conducted for experiments with each individual isolate and for the mean values for each of the two types of isolates (serotype O157 and wild-type) to determine the inactivation rates ("a" values). --------------------------------------------------------------------------- These results indicate that the E. coli O157:H7 isolates used in this study were sensitive to chlorination and were similar in resistance to that of wild-type E. coli isolates. The biocidal activity of chlorine decreases with decreasing temperature (not done in this study). The 5ºC temperature we used represents a worst-case condition for both ground water or winter surface-water temperature. A survey of disinfection practices in the United States found that water utilities maintain a median chlorine residual of 1.1 mg/L and a median exposure time of 45 minutes before the point of first use in the distribution system (10). At this level of chlorination, E. coli O157:H7 is unlikely to survive conventional water treatment practices in the United States. E. coli O157:H7 survives at a similar rate to that of wild-type E. coli in nondisinfected drinking water (11). Survival patterns and sensitivity to chlorination previously observed for the strains used in this study suggest that wild-type E. coli could serve as an adequate indicator organism for fecal contamination of water. Using wild-type E. coli to indicate E. coli O157:H7 would be useful because most analytical procedures for detecting E. coli in drinking water (e.g., assays for lactose fermentation at 44ºC to 45ºC or production of the enzyme ß-glucuronidase) cannot detect pathogenic E. coli O157:H7 strains (8). Although chlorination appears to adequately control this pathogen, not all municipal water supplies use chlorine disinfection. In addition, chlorine residual can dissipate under adverse conditions, and exposure to sunlight or organic chlorine-demand substances can greatly diminish chlorine levels. Protection of organisms associated with particulate matter, such as fecal material, can also readily decrease the biocidal activity of chlorine. These considerations are particularly important in determining the efficacy of chlorination in a recreational water setting. The results of this study indicate that the isolates studied were sensitive to chlorination. Evaluation of other isolates under differing environmental conditions would be worthy of further consideration. Acknowledgment We thank Dr. Robert V. Tauxe for his encouragement and guidance regarding this project. Dr. Rice is a microbiologist in the Microbial Contaminants Control Branch, Water Supply and Water Resources Division, National Risk Management Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio. His research focuses on detection and inactivation of waterborne pathogens and microbial indicator organisms. Address for correspondence: Eugene W. Rice, U.S. Environmental Protection Agency, 26 West M.L. King Dr., Cincinnati, OH 45268, USA; fax: 513-569-7328; e-mail: rice.gene@epa.gov. References 1. Olsen J, Miller G, Breuer T, Kennedy M, Higgins C, McGee G, et al. A waterborne outbreak of E. coli O157:H7 infections: evidence for acquired immunity. In: Program and Abstracts of the 36th Annual Meeting of Infectious Diseases Society of America, Denver, Colorado; 1998 Nov 12-15; [abstract 782]. Alexandria (VA): Infectious Disease Society of America; 1998. p. 62. 2. Blake P. Escherichia coli O157:H7 outbreak among visitors to a water park. In: Program and Abstracts of the 36th Annual Meeting Infectious Diseases Society of America, Denver, Colorado; 1998 Nov 12-15; [abstract 537]. Alexandria (VA): Infectious Diseases Society of America; 1998. p. 178. 3. Swerdlow DL, Woodruff BA, Brady RC, Griffin PM, Tippen S, Donnell HD, et al. A waterborne outbreak in Missouri of Escherichia coli O157:H7 associated with bloody diarrhea and death. Ann Int Med 1992;117:812-19. 4. Dev VJ, Main M, Gould I. Waterborne outbreak of Escherichia coli O157. Lancet 1991;337:412. 5. Ackman D, Marks S, Mack P, Caldwell M, Root T, Birkhead G. Swimming-associated hemorrhagic colitis due to Escherichia coli O157:H7 infection: evidence of prolonged contamination of a fresh water lake. Epidemiol Infect 1997;119:1-8. 6. Brewster DH, Brown MI, Robertson D, Houghton GL, Bimson J, Sharp JCM. An outbreak of Escherichia coli O157 associated with a children's paddling pool. Epidemiol Infect 1994;112:441-7. 7. Keene WE, McAnulty JM, Hoesly FC, Williams LP Jr, Hedberg K, Oxman GL, et al. A swimming-associated outbreak of hemorrhagic colitis caused by Escherichia coli O157:H7 and Shigella sonnei. N Engl J Med 1994;331:579-84. 8. Rice EW, Johnson CH, Reasoner DJ. Detection of Escherichia coli O157:H7 in water from coliform enrichment cultures. Lett Appl Microbiol 1996;23:179-82. 9. American Public Health Association. Standard methods for the examination of water and wastewater. 19th ed. Washington: The Association; 1995. 10. Water Quality Disinfection Committee. Survey of water utility disinfection practices. J Am Water Works Assoc 1992;84:121-8. 11. Rice EW, Johnson CH, Wild DK, Reasoner DJ. Survival of Escherichia coli O157:H7 in drinking water associated with a waterborne disease outbreak of hemorrhagic colitis. Lett Appl Microbiol 1992;15:38-40. -------------------------------------------------------------------------- Dispatches Fulminant Meningococcal Supraglottitis: An Emerging Infectious Syndrome? Eric Schwam*† and Jeffrey Cox† *Sturdy Memorial Hospital, Attleboro, Massachusetts, USA; and †Rhode Island Hospital, Providence, Rhode Island, USA --------------------------------------------------------------------------- We report a case of fulminant supraglottitis with dramatic external cervical swelling due to associated cellulitis. Blood cultures were positive for Neisseria meningitidis. The patient recovered completely after emergency fiberoptic intubation and appropriate antibiotic therapy. We summarize five other cases of meningococcal supraglottitis, all reported since 1995, and discuss possible pathophysiologic mechanisms. Neisseria meningitidis, a gram-negative diplococcus, can cause a broad spectrum of clinical manifestations including acute meningitis, meningococcemia, occult bacteremia, meningoencephalitis, pneumonia, conjunctivitis, dermatitis-arthritis syndrome, and urethritis (1). We describe a rare case of meningococcal supraglottitis (2-6) further complicated by cervical cellulitis. Reports of five other cases suggest an emerging clinical syndrome due to this pathogen. Case Report In January 1998, a 44-year-old woman became ill with rhinitis and sore throat, which progressed to dysphagia, dyspnea, and neck swelling during the night, requiring her to sleep sitting upright. The following morning, she went to a community hospital emergency room. The patient was alert but appeared toxic and had inspiratory stridor and a muffled voice. She had routine dental cleaning 1 week before onset of symptoms. She did not have a toothache and said she did not use tobacco or alcohol. She had had no history of splenectomy or immune deficiency. Her temperature was 38.1°C, pulse 138, cuff blood pressure 120/74 mm Hg, and respiratory rate 24. Room air pulse oximetry was 94%. The tongue was large, but not edematous, and there was neither drooling nor sublingual swelling. The soft palate and posterior pharynx, seen only with difficulty, were diffusely swollen, protruding anteriorly, and covered with exudate. Massive external swelling, tenderness, and erythema of the anterior neck extended from the chin caudad to the midsternum, obliterating all cervical landmarks. The patient did not have meningismus, crepitus, or jugular venous distention. The lungs and heart were normal. No rash was present. A portable lateral radiograph of the neck showed diffuse soft tissue cervical and epiglottic swelling with a classic "thumb sign." White blood count was 21.9 ×10(sup 9)/L with 0.70 polymorphonuclear leukocytes and 0.17 bands. Platelet count was 242 ×10(sup 9)/L. Hematocrit, electrolytes, glucose, and urea nitrogen were normal. The patient was treated with oxygen and nebulized epinephrine, ceftriaxone 1.0 gram intravenously (i.v.), and clindamycin 600 mg i.v. She was taken to the operating room, where with surgical standby, she was orally intubated with a 6.0 mm-endotracheal tube over a fiberoptic laryngoscope. The patient was transferred to a tertiary care hospital, where the antibiotics were changed to ampicillin-sulbactam, clindamycin, and gentamicin. Computed tomography (CT) scan of the neck and chest showed extensive soft tissue swelling from the oropharynx to the supraglottic region with obliteration of the airway surrounding the endotracheal tube. The adjacent parapharyngeal and cervical soft tissues down to the upper chest were also involved, but no discrete abscess was identified. Bilateral pulmonary infiltrates, atelectasis, and small pleural effusions were present, but the mediastinum was normal. The next day, two blood cultures drawn before administration of antibiotics grew N. meningitidis, serogroup Y (confirmed by the Massachusetts Department of Public Health Laboratory). The organism was sensitive to penicillin and ceftriaxone and resistant to tetracycline by the Kirby-Bauer method. Sputum culture done by endotracheal tube (after the start of antibiotics) was negative. Close family contacts and the community hospital staff members involved with airway procedures were given prophylactic antibiotics, according to guidelines (rifampin 600 mg PO bid x 4 doses, ciprofloxacin 500 mg PO x 1 dose, or ceftriaxone 250 mg intramuscularly x 1 dose) (7). Clindamycin and gentamicin were discontinued. Repeat CT scan on hospital day 5 was unchanged, but by day 7, the swelling had resolved, and the patient was extubated in the operating room. She recovered fully and was discharged on hospital day 9; she received i.v. ampicillin-sulbactam at home for an additional 7 days. She remained well, as determined by telephone contact 13 months later. Complement testing (CH-50) 13 months later was normal. Conclusions The diagnosis of supraglottitis is most consistent with the patient's clinical picture of sore throat, dysphagia, fever, muffled voice, and swollen supraglottic tissues, as seen on plain films, fiberoptic laryngoscopy, and cervical CT scan (8). However, supraglottitis is uncommonly accompanied by such dramatic external cervical swelling. Ludwig's angina was initially considered, but this diagnosis was unlikely because odontogenic infection and involvement of the sublingual or submandibular spaces were lacking. Cervical abscesses can complicate supraglottitis (9), but no abscess was detected on repeated CT scans. Necrotizing fasciitis was a possibility, especially since the erythema and swelling overlying the chest suggested mediastinal extension. In the one reported case of supraglottitis with cervical necrotizing fasciitis (10), fascial gas was demonstrated on CT scan. Surgical drainage was required for cure; anaerobic organisms were the likely cause. Furthermore, N. meningitidis, which rarely causes cellulitis, has not been reported to cause necrotizing fasciitis. The pathologic process in bacterial supraglottitis is cellulitis of the epiglottis and the surrounding upper airway. In this patient, the cellulitis was so aggressive that it extended posteriorly to the spine, antero-laterally to the cervical skin, cephalad to the pharynx, and caudad to the chest. Cultures of the epiglottis and pharynx were not performed, so the meningococcal bacteremia may have reflected secondary or coincident infection, not supraglottic infection. However, even when laryngeal cultures are performed, organisms isolated from the blood and the throat correlate poorly (11). Although the CT lung scan demonstrated pulmonary infiltrates, bacteremic meningococcal pneumonia complicating supraglottitis due to another organism is improbable. A search of Medline (the National Library of Medicine's bibliographic database of biomedical journals) from 1966 through mid-1999 located five reports of meningococcal supraglottitis: two from Colorado (2,4), one from Ohio (3), one from Singapore (5), and one from Helsinki, Finland (6). Including our patient in the series of six, the ages of the patients (three were women) were 44, 54, 60, 65, 81, and 95. Two patients had type 2 diabetes mellitus, but the others were otherwise healthy; all had fever, sore throat, and evidence of upper airway compromise. No upper airway cultures were reported, but blood cultures in all six cases were positive for N. meningitidis: serogroup B (two), serogroup Y (three), and unreported serogroup (one). No clinical evidence of meningitis or meningococcemia was found in any of the cases. Only our patient had external cervical cellulitis. All patients recovered with appropriate antibiotic treatment. The possibility of complement deficiency was investigated and ruled out in two of two patients. Two patients were treated with steroids i.v. Five of the six patients required airway intervention (three, intubation; two, urgent tracheostomy), a much higher proportion than that in a population-based case series of adults with supraglottitis (11). Thus, bacteremic meningococcal supraglottitis appears to be fulminant and life-threatening. Most meningococcal disease in the United States is sporadic. One third of all cases occur in adults, who are usually immunocompromised (e.g., by complement deficiency, corticosteroid use, or HIV infection). In one population-based study, more than half the adults had neither rash nor meningitis. Pneumonia, sinusitis, and tracheobronchitis were the main sources of bacteremic meningococcal disease. Supraglottitis was not observed (12). In particular, the proportion of meningococcal infections due to serogroup Y has been increasing nationally in the last several years. Serogroup Y is frequently associated with meningococcal pneumonia in both civilian (13) and military populations (14). Why N. meningitidis is not a more frequent cause of supraglottitis is not known. The organism is a common colonizer of the upper airways of healthy persons, and the pharynx is the suspected portal of entry of invasive and disseminated disease (1). The organism may also cause simple pharyngitis (15). The pathogenic determinants of both meningococcal disease and supraglottitis are complex and largely undefined, so we can only speculate how meningococci might cause this syndrome. Meningococcemic syndromes seem to require both epithelial and endothelial invasiveness so that the organism can cross the nasopharyngeal mucosal barrier, enter the bloodstream, and invade other blood vessel walls to produce the characteristic vasculitic organ damage. In contrast, meningococcal isolates from supraglottic syndromes seem to have relatively greater epithelial invasiveness, a propensity for contiguous local inflammatory spread, and decreased tropism for endothelial cells; these characteristics result in a more locally aggressive but less disseminated disease. The presence of various surface-expressed virulence factors (e.g., capsule, pili, cell surface proteins, and lipo-oligosaccharides) that mediate the organism's interaction with certain host cells may explain these differing pathophysiologic properties. For example, certain Opa cell surface proteins facilitate invasion of epithelial cells, while Opc proteins are more efficient at promoting invasion of endothelial cells (16). N. meningitidis can cause an inflammatory conjunctivitis that progresses to septicemia in approximately 10% of cases (17). In an analogous manner, it can (rarely) extend locally to produce a periorbital cellulitis. In one such case, isolates from the blood and periorbital aspirate of the same patient were identical except for their expression of Opa proteins and their lipo-oligosaccharide phenotype (18). Host factors (e.g., specific immune system deficiencies) could also be responsible for different disease manifestations. The most well-known example is the susceptibility of patients with terminal complement component deficiencies to neisserial infections. However, these patients have recurrent, but typical meningococcemia, so this deficiency would not be expected to contribute to the supraglottitis syndrome (19). While N. meningitidis has been known for nearly 2 centuries and blood cultures have been routinely available for decades, meningococcal supraglottitis had not been reported until 1995 (2). In contrast, incidence of supraglottitis due to Haemophilus influenzae has remained constant in adults 18 years of age or older (11). Furthermore, while one case was reported each year from 1995 to 1997, three cases have been reported in 1998-1999, from three continents, suggesting the emergence of a new meningococcal syndrome worldwide. Surveillance is needed to determine if meningococcal supraglottitis will become more than just a rarity. --------------------------------------------------------------------------- Acknowledgment We thank Dr. Gregory Jay for review of the manuscript. Dr. Schwam practices emergency medicine at Sturdy Memorial Hospital in Attleboro, Massachusetts. He is clinical instructor in medicine at Brown University Medical School and instructor in emergency medicine at the University of Massachusetts Medical School. Dr. Cox practices emergency medicine at Rhode Island Hospital in Providence, Rhode Island. He is assistant clinical professor of surgery at Brown University Medical School. Address for correspondence: Eric Schwam, Emergency Care Center, Sturdy Memorial Hospital, 211 Park Street, Attleboro, MA 02703-0963, USA; fax: 508-236-7043; e-mail: schwam@massmed.org. References 1. Apicella MA. Neisseria meningitis. In: Mandell GL, Bennett JE, Dolin R, editors. Mandell, Douglas and Bennett's principles and practice of infectious diseases. 4th ed. New York: Churchill Livingston Inc.; 1995. p. 1896. 2. Crausman RS, Jennings CA, Pluss WT. Acute epiglottitis in the adult caused by Neisseria meningitidis. Scand J Infect Dis 1995;27:77-8. 3. Nelson K, Watkins DA, Watanakunakorn C. Acute epiglottitis due to serogroup Y Neisseria meningitidis in an adult. Clin Infect Dis 1996;23:1192-3. 4. Donnelly TJ, Crausman RS: Acute supraglottitis: when a sore throat becomes severe. Geriatrics 1997;52:65-6, 69. 5. Sivalingam P, Tully AM. Acute meningococcal epiglottitis and septicaemia in a 65-year-old man. Scand J Infect Dis 1998;30:198-200. 6. Mattila PS, Carlson P. Pharyngolaryngitis caused by Neisseria meningitidis. Scand J Infect Dis 1998;30:200-1. 7. Jafari HS, Perkins BA, Wenger JD. Control and prevention of meningococcal disease. MMWR Morb Mortal Wkly Rep 1997;46:RR-5. 8. Frantz TD, Rasgon BM, Quesenberry CP Jr. Acute epiglottitis in adults: analysis of 129 cases. JAMA 1994;272:1358-60. 9. Hebert PC, Ducic Y, Boisvert D, Lamothe A. Adult epiglottitis in a Canadian setting. Laryngoscope 1998;108:64-9. 10. Nguyen R, Leclerc J. Cervical necrotizing fasciitis as a complication of acute epiglottitis. J Otolaryngol 1997;26:129-31. 11. Mayo-Smith MF, Spinale JW, Donskey CJ, Yukawa M, Li RH, Schiffman FJ. Acute epiglottitis: an 18-year experience in Rhode Island. Chest 1995;108:1640-7. 12. Stephens DS, Hajjeh RA, Baughman WS, Harvey RC, Wenger JD, Farley MM. Sporadic meningococcal disease in adults: results of a 5-year population-based study. Ann Intern Med 1995;123:937-40. 13. Racoosin J, Diaz PS, Samala U. Serogroup Y meningococcal disease—Illinois, Connecticut, and selected areas, United States, 1989-1996. MMWR Morb Mortal Wkly Rep 1996;45:1010-4. 14. Koppes GM, Ellenbogen C, Gebhart RJ. Group Y meningococcal disease in United States Air Force recruits. Am J Med 1977;62:661-6. 15. Pether JVS, Scott RJD, Hancock P. Do meningococci cause sore throats? Lancet 1994;344:1636. 16. Nassif X, Magdalene SO. Interaction of pathogenic Neisseriae with nonphagocytic cells. Clin Microbiol Rev 1995;8:376-88. 17. Moraga FA, Domingo P, Barquet N, Glasser I, Gallart A. Invasive meningococcal conjunctivitis. JAMA 1990;264:333-4. 18. Patrick CC, Glen TF, Edwards M, Estabrook M, Blake MS, Baker CJ. Variation in phenotypic expression of the Opa outer membrane protein and lipooligosaccharide of Neisseria meningitidis serogroup C causing periorbital cellulitis and bacteremia. Clin Infect Dis 1993;16:523-27. 19. Figueroa J, Andreoni J, Densen P. Complement deficiency states and meningococcal disease. Immunol Res 1993;12:295-311. -------------------------------------------------------------------------- Dispatches Genetic Evidence of Dobrava Virus in Apodemus agrarius in Hungary Jerrold J. Scharninghausen,* Hermann Meyer,† Martin Pfeffer,‡ Donald S. Davis,* and Rodney L. Honeycutt* *Texas A&M University, College Station, Texas, USA; †Federal Armed Forces Medical Academy, Munich, Germany; and ‡Ludwig Maximilians University, Munich, Germany --------------------------------------------------------------------------- Using nested polymerase chain reaction, we sequenced Dobrava virus (DOB) from the rodent Apodemus agrarius in Hungary. The samples we isolated group with DOB samples previously isolated from A. flavicollis. This grouping may indicate host switching. Hantaviruses, the causative agents of hemorrhagic fever with renal syndrome and hantavirus pulmonary syndrome, are serologically related viruses of the family Bunyaviridae and have a worldwide distribution. Unlike other bunyaviruses, hantaviruses are not transmitted by arthropod vectors. The virus is excreted in the saliva, urine, and feces of infected rodents. Humans become infected by inhalation of aerosols of dried excreta, inoculation through the conjunctiva, or entry through broken skin (1). Each viral species within the genus Hantavirus is primarily associated with a single rodent species, although accidental infections have been reported in other mammals (2). Four primary reservoirs for hantaviruses are found in Europe: Rattus norvegicus, Seoul virus; Clethrionomys glareolus, Puumala virus (PUU); Apodemus flavicollis, Dobrava virus (DOB); and Microtus arvalis, Tula virus (TUL) (3). An additional rodent species, Apodemus agrarius, is the primary reservoir of Hantaan virus (HTN), the causative agent of Korean hemorrhagic fever, which is found throughout Korea and China but not in Europe. Recently, DOB was isolated from A. agrarius in Estonia (4) and Russia (5). Hantaviruses are present in Hungary; however, their particular virus strains or species and their distribution are unknown. PUU has been confirmed in western Hungary, but the species in eastern Hungary have not been determined (6). During field surveys of indigenous rodents in areas used by NATO forces between 1995 and 1996, nine rodents were collected at Tazar Air Force Base, outside Kaposvar, eastern Hungary. A total of 210 trap nights resulted in a trap return of 4.28%. Two A. agrarius, four Microtus agrestis, and three Mus musculus were captured. Animals were live-trapped and euthanized with halazone. Tissue was collected from lungs and kidneys; urine, if present, was also collected. Total RNA was extracted from collected tissues by using the FastRNA KitGreen (Bio 101), according to the supplier's recommendations. RNA was converted to cDNA and amplified by the Titan One Tube RT-PCR System (Boehringer, Mannheim, Germany), according to the manufacturer's recommendations. Degenerate primers M-4 (ATGAARGCNGAWGA RNTNACMCCNGG) and M-9 (TGRYCNAGYTG TATYCCCATWGATTG) were used to amplify a 583-bp section corresponding to nt positions 605 to 1,188 in the S segment of HTN strain X95077. No DOB strains were used in the laboratory. A target sequence of 397 bp (nt from 692 to 1,089) was amplified by nested polymerase chain reaction, and both strands of the amplicons were sequenced. Amplified bands of the predicted size were demonstrable in all tissue samples investigated from the two specimens of A. agrarius (designated Tazar-2, Tazar-8). All other samples tested negative. Sequence validity was confirmed by sequencing amplified lung and kidney products from each infected animal. The respective sequences from Tazar-2 and Tazar-8 differed in 3 of the 321 nucleotides examined (99% identity). Both sequences were aligned with the corresponding S-segment section of several DOBs and five other European and Asian hantaviruses. The nucleotide sequence identities between Tazar virus (found in this study) and related viral lineages (Figure) included Russian DOB from A. agrarius, 88%; Estonian DOB from A. agrarius, 86%-87%; Bosnia DOB from A. flavicollis, 88%; Greek DOB from A. flavicollis, 88%-85%; Sapporo rat virus, 70%; HTN, 68%; Khabarousk, 57%; PUU, 56%; TUL, 53%; and Sin Nombre, 48%. These results indicate that, on the basis of sequence similarity, both Tazar samples are hantaviruses most closely related to DOB. Although the sequence data are limited, phylogenetic analyses linked Tazar-2 and Tazar-8 isolated from A. agrarius to a group of DOB previously isolated from A. flavicollis. The A. agrarius DOB from Russia (5) did not support monophyly (common ancestry) for DOB isolated from A. agrarius populations in Hungary and Russia. Other representative hantaviruses, including PUU, TUL, HTN, and Sapporo rat virus, were more basal in the phylogeny (Figure). HTN infects A. agrarius populations in Asia but has not been isolated in Europe. Direct enzyme-linked immunosorbent assay has demonstrated the presence of Hantaan-like antigens in A. agrarius in the former republic of Czechoslovakia (5.5%, [7]), the European regions of the former Soviet Union (5.3%, [8]) (28.5%, [9]), and Serbia (2.2%, [10]). Because HTN sequences have not been reported in Europe and HTN and DOB have similar immunologic responses, the earlier findings of Hantaan-like antigens in Europe may represent a more widespread occurrence of DOB in European populations of A. agrarius. Recent verification of DOB in populations of A. agrarius in Estonia (4), Russia (5), and now Hungary supports this conclusion. Figure. Cladogram derived from nucleotide sequences of Tazar-2, Tazar-8, and other hantaviruses. Numbers denote Genbank accession numbers. The cladogram was derived from the neighbor-joining estimated phylogeny and bootstrap analysis using p-distance estimates. The phylogenetic [Fig] analysis was performed by using PAUP 3.1.1 (vers. 4.0.0d64). Numbers at each internode or bifurcation represent bootstrap support based on 1,000 replicates. A Sin Nombre virus sequence (L37904) was used as the outgroup to root the tree. Tazar 2 and Tazar 8 have been submitted to Genbank and received accession numbers AF085336 and AF085337, respectively. The occurrence of DOB in both A. agrarius and A. flavicollis provides an opportunity to evaluate the hypothesis concerning distribution of hantaviruses in related rodent hosts. Phylogenetically different Sin Nombre–like viruses have been found in different populations within species of peromyscine rodents that vary ecologically and geographically throughout their range (11). Although these authors found evidence of cospeciation between the rodent host phylogeny and the host-borne hantavirus phylogeny, evidence of host switching was observed with Peromyscus leucopus–borne New York virus grouped with P. maniculatus–borne viruses rather than with other P. leucopus–borne viruses. Our phylogenetic analysis (Figure) indicates a closer relationship between the A. agrarius DOB from Hungary and A. flavicollis DOB, with Russian/Estonian DOB representing a sister-group to this clade. This lack of monophyly for the A. agrarius DOB may suggest host switching between A. agrarius and A. flavicollis similar to that between P. leucopus and P. maniculatus. Nevertheless, the relationships between the isolated DOB lineages suggest a more basal position for the A. agrarius DOB than for A. flavicollis DOB lineages. A. flavicollis ranges throughout much of Western Europe eastward to the Ural Mountains, and A. agrarius ranges from Eastern Europe eastward to the Pacific Ocean, covering most of the Asian continent (12). Given the extensive range of both species, an examination of other populations within each species, as well as other species of Apodemus, might allow correlation of the viral and rodent host phylogeny. The pattern of divergence for hantaviruses in New World peromyscine rodent species may be mirrored in Old World arvicoline rodents. Therefore, the existence of more than one hantavirus in A. agrarius may reflect geographic variation within the species or host switching in regions where two host species are potentially sympatric. --------------------------------------------------------------------------- Dr. Scharinghausen, an active-duty Army captain, is completing his Ph.D. at Texas A&M University in Wildlife and Fisheries Sciences. His area of expertise is mammalogy, and his research interests include zoonoses and molecular biology. Address for correspondence: Jerrold J. Scharninghausen, Department of Wildlife & Fisheries Sciences, Texas A&M University, College Station, Texas 77843, USA; e-mail: 100304.2713@compuserve.com. References 1. Hjelle B, Jenison S, Goade D, Green W, Feddersen R, Scott A. Hantaviruses: clinical, microbiologic, and epidemiologic aspects. Crit Rev Clin Lab Sci 1995;32:469-508. 2. Childs J, Glass G, Korch G, LeDuc J. Prospective seroepidemiology of hantaviruses and population dynamics of small mammal communities of Baltimore, Maryland. Am J Trop Med Hyg 1987;37:648-62. 3. Plyusnin A, Cheng Y, Vapalahti O, Pejcoch M, Unar J, Jelinkova Z, et al. Genetic variation in Tula hantaviruses: sequence analysis of the S and M segments of strains from Central Europe. Virus Res 1995;39:237-50. 4. Nemirov K, Vapalahti O, Lundkvist A, Vasilen Golovljova I, Plyusnina A, Niemmimaa J, et al. Isolation and characterization of Dobrava hantavirus in the striped field mouse (Apodemus agrarius) in Estonia. J Gen Virol 1999;80:371-9. 5. Plyusnin A, Nemirov K, Apekina N, Plyusnina A, Lundkvist A, Vaheri A. Dobrava hantavirus in Russia. Lancet 1999;353:207. 6. Centers for Disease Control and Prevention. The Fourth International Conference on HFRS and Hantaviruses; 1998 March 5-7; Atlanta, Georgia. Atlanta: U.S. Department of Health and Human Services; 1998. 7. Danes L, Tkachenko E, Ivanov A, Lim D, Rezapkin G, Dzagurova T. Hemorrhagic fever with renal syndrome in Czechoslovakia: detection of antigen in small terrestrial mammals and specific serum antibodies in man. Journal of Hygiene, Epidemiology, Microbiology, and Immunology 1986;30:79-85. 8. Tkachenko E, Ivanov A, Donets M, Miasnikov Y, Ryltseva E, Gaponova L, et al. Potential reservoir and vectors of haemorrhagic fever with renal syndrome (HFRS) in the USSR. Annales de Société Belgique Medecíne Tropique 1983;63:267-9. 9. Gavrilovskaya I, Apekina N, Myasnikov Y, Brenshtein A, Ryltseva E, Gorbachkova E, et al. Features of circulation of hemorrhagic fever with renal syndrome (HFRS) virus among small mammals in the European USSR. Arch Virol 1983;75:313-6. 10. Gligic A, Obradovic M, Stojanovic R, Hlaca D, Antonijevic B, Arnautovic A, et al. Hemorrhagic fever with renal syndrome in Yugoslavia: detection of hantaviral antigen and antibodies in wild rodents and serological diagnosis of human disease. Scand J Infect Dis 1988;20:261-6. 11. Morzunov S, Rowe J, Ksiazek G, Peters CJ, St Jeor S, Nichol S. Genetic analysis of the diversity and origin of hantavirus in Peromyscus leucopus mice in North America. J Virol 1998;1:57-64. 12. Nowak RM, editor. Walker's mammals of the world. 5th ed. Baltimore: Johns Hopkins University Press; 1991. -------------------------------------------------------------------------- Dispatches Bacterial Resistance to Ciprofloxacin in Greece: Results from the National Electronic Surveillance System A.C. Vatopoulos, V. Kalapothaki, Greek Network for the Surveillance of Antimicrobial Resistance (ft1), and N.J. Legakis Athens University, Athens (Goudi), Greece --------------------------------------------------------------------------- According to 1997 susceptibility data from the National Electronic System for the Surveillance of Antimicrobial Resistance, Greece has high rates of ciprofloxacin resistance. For most species, the frequency of ciprofloxacin- resistant isolates (from highest to lowest, by patient setting) was as follows: intensive care unit > surgical > medical > outpatient. Most ciprofloxacin-resistant strains were multidrug resistant. Soon after the broad-spectrum, highly effective antibiotics fluoroquinolones were introduced, their extensive use and misuse in hospitals and communities, as well as in veterinary medicine, have led to the emergence and spread of resistant strains (1,2). Highly divergent rates of fluoroquinolone resistance in both community-acquired and nosocomial pathogens have been reported worldwide (2). Many factors, including patient characteristics, local epidemiologic factors, antibiotic policies, over-the-counter use (which often leads to inadequate use), lower standard of living in developing countries, lack of information on the prudent use of antibiotics, and use of antibiotics in animal husbandry may contribute to the emergence of quinolone-resistant organisms. Surveillance is an integral part of controlling resistance, and local and national surveys to identify, monitor, and study the epidemiology of the emergence and spread of resistant isolates are needed (3). To identify national trends and local differences in the epidemiology of quinolone resistance in Greece, we report 1997 ciprofloxacin susceptibility data from the National Electronic System for the Surveillance of Antimicrobial Resistance. The National Electronic System for the Surveillance of Antimicrobial Resistance was introduced in Greece 3 years ago. Involving 17 hospitals throughout Greece, the system analyzes the routine results of the antibiotic sensitivity tests performed in hospital microbiology laboratories by using WHONET software (4). In our analysis we included 11,097 isolates (4,204 from medical wards, 2,897 from surgical wards, 1,724 from intensive care units [ICU], and 2,272 from outpatient departments) (Table 1). We focused on the bacteria most frequently encountered in Greek hospitals (National Electronic System for the Surveillance of Antimicrobial Resistance [www.mednet.gr/whonet]; N.J. Legakis, Enare Sentry, unpub. data): Escherichia coli, Klebsiella pneumoniae, Enterobacter species, Pseudomonas aeruginosa, Acinetobacter baumanii, and Staphylococcus aureus. These species are also the most important nosocomial pathogens in most parts of the world in terms of rate of isolation, pathogenicity, and virulence (5,6). Isolation Table 1. Isolates included in the analysis(sup a) and identification------------------------------------------------------------- were Type of ward performed by ------------------------------------------- standard Medi- Surgi- Outpa- methods at Species cal cal ICU tients All the (sup b) microbiology ------------------------------------------------------------- laboratories Escherichia 2,100 1,114 94 1,571 4,879 of each coli hospital Pseudomonas 672 527 570 195 1,964 participating aeruginosa in the Staphylococcus 452 467 318 248 1,485 network. The aureus susceptibility testing Enterobacter 396 332 198 142 1,068 methods were spp. Kirby-Bauer Klebsiella 419 224 177 96 916 disk pneumoniae diffusion (7 Acinetobacter 165 233 367 20 785 hospitals); spp. Sensititre All 4,204 2,897 1,724 2,272 11,097 (Sensititre, Salem, NH) ------------------------------------------------------------- (1); Pasco (sup a)One isolate per species per patient (the first isolated) (Difco, is shown. bICU, intensive care unit. Detroit, MI) (8); and VITEK (Bieux-Merieux Marcy l'Etoile, France) (1). The actual zone diameters or MICs (not the interpretations of the tests) were entered into WHONET. The chi-square test was used to evaluate differences in resistance rates between types of wards, as well as between clinical specimens. Pearson's correlation coefficients were calculated for possible associations between resistance rates and hospital size. The resistance rate to ciprofloxacin by type of ward, clinical specimen, and bacterial species is shown in Table 2. There is a stepwise decrease in the frequency of isolation of ciprofloxacin-resistant isolates (ciprofloxacin resistance in isolates from ICU patients > isolates from surgical patients > isolates from medical patients > isolates from outpatients). These differences were significant (p <0.01), with the exception of decreases in resistance rates for E. coli between surgical wards and ICUs; for Enterobacter spp. between medical and surgical wards; for Acinetobacter spp. between outpatients, medical, and surgical wards; and for S. aureus between medical and surgical wards. Moreover, for P. aeruginosa, the resistance rates were significantly higher in medical than in surgical wards (p = 0.00097). As for clinical specimens, each bacterial species followed a different pattern (Table 2). In medical wards, enterobacterial strains isolated from purulent infections were more often resistant to ciprofloxacin, but this difference was statistically significant only for K. pneumoniae (p = 0.012). In surgical wards, blood and respiratory isolates were more often resistant, but this difference was significant only for Enterobacter spp. (p = 0.02). On the other hand, ciprofloxacin-resistant P. aeruginosa strains were more frequently isolated (p = 0.0021) in medical wards from urine and in surgical wards from urine and blood as opposed to all other specimens (p = 0.0005). No significant differences were observed in the rate of isolation of ciprofloxacin-resistant A. baumanii strains among the various clinical specimens. S. aureus strains resistant to ciprofloxacin were mostly methicillin-resistant (MRSA) (Table 2). Very low resistance rates were observed in P. aeruginosa isolated from ear infections, especially from outpatients. Table 2. Ciprofloxacin resistance by specimen and type of ward(sup a) ------------------------------------------------------------------------- Outpatients Medical Surgical ICU ------------- ------------ ------------- ---------- No. %R No. %R No. %R No. %R (sup b) ------------------------------------------------------------------------- Escherichia coli Urine 1,191 5.0 1,572 5.5 597 8.5 39 10.2 Blood - 195 6.9 14 18.1 5 0.0 Respiratory - 56 2.1 - 23 9.0 Pus - 33 12.1 203 8.4 11 27.8 Other 380 4.5 244 7.5 300 6.5 16 20.0 All 1,571 3.7 2,100 5.6 1,114 8.2 94 13.3 Salmonella spp. Stool 195 0.7 Klebsiella pneumoniae Urine 62 6.6 254 15.5 85 19.8 28 64.0 Blood - 45 11.3 10 9.8 18 72.3 Respiratory - 62 9.8 12 50.0 90 69.8 Pus - 14 50.0 42 19.0 0 0.0 Other 34 3.1 44 18.5 79 28.3 41 65.4 All 96 5.4 419 15.8 226 23.9 177 67.7 Serratia marcences All 76 7.7 20 45.2 (sup c)(sup c) Enterobacter spp. Urine 76 12.0 190 29.7 85 32.0 24 75.4 Blood - 37 21.8 13 54.2 24 66.6 Respiratory - 76 6.3 10 40.2 58 48.6 Pus - 22 36.8 138 18.5 27 67.6 Other 66 10.8 71 16.9 86 23.3 65 69.0 All 142 11.6 396 22.2 332 24.8 198 62.2 Pseudomonas aeruginosa Urine 51 31.0 270 44.0 171 40.7 70 79.3 Blood 0 0.0 24 20.6 13 46.5 29 75.6 Respiratory 11 18.2 258 34.4 29 44.6 379 62.9 Pus 18 11.3 35 31.6 147 22.6 16 69.5 Ear 72 1.7 7 47.3 30 3.7 0 0.0 Other 43 18.8 78 26.9 137 25.9 76 66.9 All 195 16.7 672 37.5 527 28.2 570 66.4 Acinetobacter spp. Urine - 72 62.6 32 65.9 34 94.4 Blood - 18 38.7 16 69.0 40 92.3 Respiratory - 38 49.7 11 100.0 190 91.0 Pus - 13 61.8 87 60.1 19 94.8 Other - 24 62.5 87 69.1 84 78.9 All 20 45.1 165 56.8 233 66.6 367 88.4 Staphylococcus aureus Urine - 37 32.9 16 31.0 - Blood - 101 51.0 15 67.0 40 62.7 Respiratory - 123 45.3 28 57.1 221 65.8 Pus 104 18.2 88 21.6 272 30.8 14 71.4 Ear 52 3.8 - - - Other 92 10.3 103 25.6 136 31.4 43 67.4 All 248 12.8 452 30.5 467 33.0 318 63.6 MRSAd 40 56.7 140 69.1 176 75.3 375 94.3 MSSAe 184 1.7 256 12.4 219 6.5 92 4.6 ------------------------------------------------------------------------- (sup a)One isolate per patient (the first isolated) is shown. (sup b)R, resistant. (sup c)Medical and surgical wards combined. (sup d)MRSA, methicillin-resistant S. aureus. (sup e)MSSA, methicillin-sensitive S. aureus. Approximately 75% of K. pneumoniae, 87% of Enterobacter spp., 55% of P. aeruginosa, 76% of A. baumanii, and 75% of MRSA strains were drug resistant to at least three different classes (Table 3). However, 15% of the ciprofloxacin-resistant E. coli were resistant only to this antibiotic, and 25% had additional resistance only to cotrimoxazole. Moreover, 48% of ciprofloxacin-resistant but methicillin-sensitive S. aureus were resistant only to chloramphenicol. Table 3. Resistant phenotypes of ciprofloxacin-resistant isolates to other classes of antibiotics(sup a) --------------------------------------------------------------------------- Klebsiella pneumoniae Enterobacter spp Escherichia coli ------------------- ------------------ ------------------- Phenotypeb No. % Phenotype No. % Phenotype No. % F 4 3.7 F 0 0 F 25 15.1 DBXF 9 8.4 IF 4 2.5 IDBXF 16 9.6 IDB F 16 15.0 IDB F 7 4.4 IXF 29 17.5 IDBXF 64 59.8 IDBXF 131 82.9 XF 42 25.3 all other 14 13.1 all other 16 10.1 all other 54 32.5 All 107 100.0 All 158 100.0 All 166 100.0 Pseudomonas Acinetobacter aeruginosa baumanii ------------------- ------------------ Phenotype No. % Phenotype No. % F 10 7.3 F 0 0.0 1DM F 14 10.2 SMD X 5 10.0 1DMNF 23 16.8 D XF 15 30.0 1 M F 40 29.2 MD XF 23 46.0 all other 50 36.5 all other 7 14.0 All 137 100.0 All 50 100.0 Staphylococcus aureus MRSA MSSA --------------------------------------------- Phenotype No. % Phenotype No. % F 0 0.0 F 7 10.3 OG E F 23 11.3 E F 9 13.2 OG ECF 44 21.7 CF 33 48.5 OGXECF 84 41.4 all other 52 25.6 all other 19 27.9 All 203 100.0 All 68 100.0 ------------------------------------------------------------------------ (sup a)All wards, intensive care units isolates are not included. b1, piperacillin; B, tobramycin; C, chloramphenicol; D, ceftazidime; E, erythromycin; F, ciprofloxacin; G, gentamicin; I, cefoxitin; M, amikacin; N, imipenem; O, oxacillin; S, amoxicillin/sulbactam; X, cotrimoxazole; MRSA, methicillin-resistant S. aureus; MSSA, methicillin-sensitive S. aureus. When we plotted resistance rates to ciprofloxacin against the number of beds in each hospital, we found no correlation (Figure). The rate of isolation of ciprofloxacin-resistant isolates varied greatly by hospital for all species examined: from 1% to 15% for E. coli, 1% to 23% for K. pneumoniae, 1% to 33% for Enterobacter spp., 11% to 33% for P. aeruginosa, 29% to 73% for A. baumanii, and 11% to 48% for S. aureus. Ciprofloxacin resistance was observed in hospitals throughout Greece. In Europe and North America, a striking difference in the incidence of bacterial resistance to quinolones has been observed between nosocomial and community-acquired infections; resistance is only rarely encountered among the latter (2,7). The incidence of resistance to fluoroquinolones in bacteria isolated from hospital-acquired infections varies among bacterial species, clinical settings, and countries and may be related to local epidemic spread of a few clones (2). The highest incidence of resistance is among P. aeruginosa, Acinetobacter spp., Serratia marcescens, and particularly MRSA strains (8). Our results place Greece among the countries with high resistance levels to quinolones. Although quinolones are among the antibiotics restricted by the Greek Ministry of Health and Welfare, the mean national level of quinolone resistance has increased in most bacterial species during the last 5 years (9). The 3.7% quinolone resistance rate [fig] among E. coli isolated from outpatients is almost double that in Fig. Resistance rates to other industrialized countries (2). ciprofloxavin in each hospital by This high rate may be due to the use number of beds and geographic of quinolones, and especially area of the hospital. Only norfloxacin, as a first-line hospitals with more than 20 antibiotic in Greece to treat isolates are included. uncomplicated urinary tract infections (Isolates from all wards but in the outpatient setting. Free access not intensive care units.) to fluoroquinolones has also been incriminated in increased quinolone resistance in industrialized and developing countries (10,11). The low rate of quinolone resistance in salmonellas, compared with other countries (12,13), may be due to infrequent use of quinolones in farm animals in Greece. Among Enterobacteriaceae, quinolone resistance seems to be higher in K. pneumoniae and Enterobacter spp. than in S. marcescens. The high level of resistance in ICUs was expected since ICUs are well-known focuses of antimicrobial resistance (14). Hospitalization in ICUs was an independent risk factor for acquiring infection by multidrug-resistant strains in Greece (15). Moreover, ICU patients are often colonized with endemic, multidrug-resistant strains, which often spread to other wards (16). We found higher rates of isolation of quinolone-resistant strains of some species in the surgical wards than in medical wards. Patients at high risk for a resistant nosocomial infection (e.g., cancer patients, immunosupressed patients) are usually in medical wards. High resistance in the surgical wards could be the result of nursing practices or unnecessary prophylactic administration of antibiotics, both of which should be further evaluated. Most quinolone-resistant strains in Greece are also resistant to other clinically relevant antibiotics. The possible clinical and epidemiologic importance of the newly described multidrug efflux pumps in multidrug resistance, mainly in P. aeruginosa, is under investigation worldwide (17). Moreover, the marginal susceptibility of S. aureus to quinolones and the ease with which mutations affecting susceptibility can occur in this species contribute to the observed high rates of quinolone resistance. MRSA strains are no more likely to develop resistance to quinolones than other staphylococci (8). In any case, the favorable accumulation of different traits in quinolone-resistant strains or, alternatively, the favorable potential for mutation to quinolone resistance in multidrug-resistant strains has not been proved. Epidemiologic parameters, and more specifically the sequential introduction of various antibiotic classes in most of the world and in Greek hospitals, could explain multidrug resistance. The extensive aminoglycoside and beta-lactamase use in the 1980s is responsible for the high prevalence of multidrug-resistant plasmids and transposons found in the nosocomial strains of various bacterial genera in Greek hospitals (18-20). The strains harboring these plasmids can survive in the hospital environment and become the best candidates for selection of resistant mutants under the pressure of quinolones. That quinolone-resistant strains are found in hospitals in all parts of Greece and resistance is not associated with the size of the hospital or its geographic area are consistent with the high prescription rate for quinolones. However, the isolation rate of resistant strains varied considerably by hospital, perhaps because of local epidemiologic factors (e.g., prescribing or nursing habits) or possible (epidemic) spread of strains among patients. This study has limitations. First, it is based on routine data generated in the microbiology laboratories of participating hospitals. Sometimes different antibiotics are tested in each hospital, which limits the possibility for interhospital comparisons. Moreover, different methods for susceptibility testing are used in each hospital. Data such as antibiotic consumption or days of hospitalization are not available since they are not included as information in the WHONET software and they are difficult and time-consuming to collect routinely. Quinolone use is a well-proven independent risk factor for resistance (21,22). Nevertheless, local differences indicate that other epidemiologic parameters should be further evaluated. --------------------------------------------------------------------------- The National Electronic System for the Surveillance of Antimicrobial Resistance has been supported in part by a grant from the Greek Ministry of Health and Welfare. The following hospitals participate in the system: Polycliniki General Hospital, Agia Olga General Hospital, Elpis General Hospital, First IKA Hospital of Athens, Agios Savas Cancer Hospital, Sismanoglion General Hospital, Hippocration General Hospital, Areteion University Hospital, Venizelio General Hospital, University Hospital of Alexandroupolis, University Hospital of Ioannina, General Hospital of Xanthi, Threassio General Hospital, Tzannio General Hospital, Asclepeion Voulas General Hospital, Theagenio Cancer Hospital, and Hippocration Hospital Thessaloniki. Dr. Vatopoulos is a medical microbiologist and assistant professor in the Department of Hygiene and Epidemiology, Medical School, Athens University. His chief research interest is the molecular epidemiology of antibiotic resistance in bacteria (mainly gram-negative). He is now involved in the establishment and operation of an electronic network for the surveillance of antibiotic resistance in Greece. Address for correspondence: A.C. Vatopoulos, Department of Hygiene & Epidemiology, Medical School, Athens University, 115 27 Athens (Goudi), Greece; fax: 30-1-7704225; e-mail: avatopou@cc.uoa.gr. (footnote 1)G. Antoniadis, E. Arhondidou, S. Chatzipanagiotou, E. Chinou, A. Chrysaki, V. Daniilidis, G. Genimata, H. Gessouli, P. Golemati, E. Kaili-Papadopoulou, A. Kansuzidou, D. Kailis, E. Kaitsa, M. Kanelopoulou, Sp. Kitsou-Kyriakopoulou, Z. Komninou, E. Kouskouni, Chr. Koutsia-Karouzou, S. Ktenidou-Kartali, V. Liakou, H. Malamou-Lada, H. Mercuri, C. Nicolopoulou, A. Pagkali, E. Panagiotou, E. Papafragas, A. Perogamvros, C. Poulopoulou, D. Sofianou, G. Theodoropoulou-Rodiou, S. Thermogianni, E. Trikka-Graphakos, O. Vavatsi-Manou, M. Ventouri, E. Vogiatzakis, A. Xanthaki, Chr. Zagora, E. Chatzidaki, G. Papoutsakis. References 1. Blondeau JM, Yaschuk Y, Canadian ciprofloxacin susceptibility study. Comparative study from 15 medical centers. Antimicrob Agents Chemother 1996;40:1729-32. 2. Acar JF, Goldstein FW. Trends in bacterial resistance to fluoroquinolones. Clin Infect Dis 1997;24:S67-73. 3. Report of the American Society for Microbiology Task Force on Antibiotic Resistance. Washington: American Society for Microbiology; 1995. p. 1-23. 4. Stelling JM, O'Brien TF. Surveillance of antimicrobial resistance: the WHONET program. Clin Infect Dis 1997;24:S157-68. 5. Emori TG, Gaynes RP. An overview of nosocomial infections, including the role of the microbiology laboratory. Clin Microbiol Rev 1993;6:428-42. 6. Wartz MN. Hospital-acquired infections: diseases with increasingly limited therapies. Proc Natl Acad Sci U S A 1994;91:2420-7. 7. Goldstein FW, Acar JF. Epidemiology of quinolone resistance: Europe and North and South America. Drugs 1995;49:S36-42. 8. Sanders CC, Sanders WE Jr, Thomson. Fluoroquinolone resistance in Staphylococci: new challenges. Eur J Clin Microbiol Infect Dis 1995;Suppl 1:6-11. 9. Legakis NJ, Tzouvelekis LS, Tsakris A, Legakis JN, Vatopoulos AC. On the incidence of antibiotic resistance among aerobic gram-negative rods isolated in Greek hospitals. J Hosp Infect 1993;24:233-7. 10. Kresken M, Hafner D, Mittermayer H, Verbist L, Bergogne-Berezin E, Giamarellou H, et al. Prevalence of fluoroquinolone resistance in Europe. Study Group `Bacterial Resistance' of the Paul-Ehrlich-Society for Chemotherapy. Infection 1994;22:S90-8. 11. Casellas JM, Blanco MG, Pinto ME. The sleeping giant: antimicrobial resistance. Infect Dis Clin North Am 1994;8:29-45. 12. Tassios PT, Markogiannakis A, Vatopoulos AC, Katsanikou E, Velonakis EN, Kourea-Kremastinou J, et al. Molecular epidemiology of antibiotic resistance of Salmonella enteritidis during a seven year period in Greece. J Clin Microbiol 1997;35:1316-21. 13. Tassios PT, Vatopoulos AC, Mainas E, Gennimata D, Papadakis J, Tsiftsoglou A, et al. Molecular analysis of ampicillin-resistant sporadic Salmonella typhi and Salmonella paratyphi B clinical isolates. Clinical Microbiology and Infection 1997;3:317-23. 14. Archibald L, Phillips L, Monnet D, McGrowan JE, Tenover F, Gaynes R. Antimicrobial resistance in isolates from inpatients and outpatients in the United States: increasing importance of the intensive care unit. Clin Infect Dis 1997;24:211-5. 15. Vatopoulos AC, Kalapothaki V, Legakis NJ, the Hellenic Antibiotic Resistance Study Group. Risk factors for nosocomial infections caused by gram-negative bacilli. J Hosp Infect 1996;34:11-22. 16. Tassios PT, Gennimata V, Spaliara-Kalogeropoulou L, Kairis D, Koutsia C, Vatopoulos A, et al. Multiresistant Pseudomonas aeruginosa serogroup O:11 outbreak in an intensive care unit. Clinical Microbiology and Infection 1997;3:621-8. 17. Nikaido H. Antibiotic resistance caused by gram-negative multidrug efflux pumps. Clin Infect Dis 1998;Suppl 1:S32-41. 18. Vatopoulos A, Phillipon A, Tsouvelekis L, Komninou Z, Legakis NJ. Prevalence of a transferable SHV-5 type ß-lactamase in clinical isolates of Klebsiella pneumoniae and Escherichia coli in Greece. J Antimicrob Chemother 1990;26:635-48. 19. Tsakris A, Johnson AP, George RC, Mehtar S, Vatopoulos AC. Distribution and transferability of plasmids encoding trimethoprim resistance in urinary pathogens from Greece. J Med Microbiol 1991;34:153-7. 20. Vatopoulos AC, Tsakris A, Tzouvelekis LS, Legakis NJ, Pitt TL, Miller GH, et al. Diversity of aminoglycoside resistance in Enterobacter cloacae in Greece. Eur J Clin Microbiol Infect Dis 1992;11:131-8. 21. Richard P, Delangle MH, Merrien D, Barille S, Reynaud A, Minozzi C, et al. Fluoroquinolone use and fluoroquinolone resistance: Is there an association? Clin Infect Dis 1994;19:54-9. 22. Carratala J, Fernandez-Sevilla A, Tubau F, Callis M, Gudiol F. Emergence of quinolone-resistant Escherichia coli bacteremia in neutropenic patients with cancer who have received prophylactic norfloxacin. Clin Infect Dis 1995;20:557-60. -------------------------------------------------------------------------- Dispatches Emergence of Related Nontoxigenic Corynebacterium diphtheriae Biotype mitis Strains in Western Europe Guido Funke,* Martin Altwegg,* Lars Frommelt,† and Alexander von Graevenitz* *University of Zurich, Zurich, Switzerland; and †Endo-Klinik, Hamburg, Germany --------------------------------------------------------------------------- We report on 17 isolates of Corynebacterium diphtheriae biotype mitis with related ribotypes from Switzerland, Germany, and France. Isolates came from skin and subcutaneous infections of injecting drug users, homeless persons, prisoners, and elderly orthopedic patients with joint prostheses or primary joint infections. Such isolates had only been observed in Switzerland. Nontoxigenic Corynebacterium diphtheriae strains were recovered from approximately 1 per 1,000 throat swabs from immunized British military personnel in Germany from 1993 to 1995 (1). Such nontoxigenic C. diphtheriae biotype mitis isolates had been described in skin, throat, and blood cultures of Swiss injecting drug users; 32 of the isolates belonged to the same clone (2,3). Our study demonstrates that this clone and closely related clones occurred between 1990 and 1997 in two other European countries, and only sometimes in persons with poor hygiene. Five C. diphtheriae isolates came from two laboratories in Zurich and Bern, Switzerland; 11 from four laboratories in Hamburg, Germany; and 1 from Paris, France (Table). They were identified in Zurich as C. diphtheriae biotype mitis (4); that is, they were nonlipophilic, were nitrate reductase-positive, and did not ferment glycogen. By polymerase chain reaction techniques (5), the diphtheria toxin gene was not detected in any isolate. For some isolates, the Elek test was also performed; it was consistently negative. For ribotyping, DNA was isolated, digested with either EcoRI or PvuII, electrophoresed, blotted, and probed for rDNA as described elsewhere (2,3). Disk susceptibility testing to tetracycline and MIC determinations were done according to National Committee for Clinical Laboratory Standards methods (6). All other bacteria isolated were also identified and serotyped in Zurich. Table. Nontoxigenic Corynebacterium diphtheriae isolates ----------------------------------------------------------------------------------------------------- Ribotyping pattern(sup c) ----------- Patients Clinical Isolate Date Place of sex, age diagnosis/ Other bacteria Underlying no. isolated isolation EcoRI PvuII (yr) source isolated conditions (sup a) (sup b) ----------------------------------------------------------------------------------------------------- 2012 1990 Paris A C m, 6 Endocarditis/ — Skin lesions, blood culture scabies 2410 2/1995 Hamburg A B m, 30 Ulcus/lower Staphylococcus Prisoner leg aureus, Streptococcus pyogenes 2689 3/1995 Hamburg A A f, 82 Puncture/hip S. epidermidis, Total hip joint joint Corynebacterium endoprosthesis pseudodiphtherit- icum 1682 9/1995 Hamburg A B m, 45 Wound swab S. aureus, Homelessness S. pyogenes 1935 3/1996 Hamburg A B m, 58 Wound swab S. aureus, Alcoholism, S. pyogenes homelessness 1836 3/1996 Hamburg A B m, 45 Abscess/lower S. aureus, Injecting leg S. pyogenes drug use 2661 3/1996 Hamburg A A m, 35 Arthroscopy Peptostreptococcus Osteosynthesis, aspirate/knee joint prevotii,P. magnus arthrotomy, knee joint after trauma 2658 4/1996 Hamburg A A f, 76 Fistula/hip S. aureus Total hip joint joint endoprosthesis 2674 5/1996 Hamburg A A f, 62 Aspirate/hip Streptococcus Total hip joint joint group C endoprosthesis 2670 7/1996 Hamburg A A f, 81 Aspirate/knee — Total knee joint joint endoprosthesis 2667 7/1996 Hamburg A A m, 23 Swab/hip joint Coagulase-negative Total hip joint staphylococci endoprosthesis 2413 11/1996 Hamburg A B m, 36 Ulcus/lower S. aureus, Prisoner leg Streptococcus group C/G 2446 12/1996 Zurich A A f, 38 Wound/upper S. aureus, S. Injecting drug leg pyogenes use 2464 1/1997 Zurich A A m, 39 Ulcera/lower Escherichia coli, Injecting drug arm use S. aureus, S. pyogenes 2473 1/1997 Zurich A A m, 38 Ulcera/leg E. coli, Injecting drug Streptococcus use group C, G 2475 1/1997 Zurich A A m, 31 Ulcus/pretibialPseudomonas Injecting drug aeruginosa, E. use coli, Streptococcus group C, G 480 5/1997 Bern A A m, 39 Ulcus/upper legS. aureus, Injecting Clostridium drug use perfringens, mixed anaerobic flora ----------------------------------------------------------------------------------------------------- (sup a)Month/Year (sup b)Paris, France; Hamburg, Germany; Zurich, Switzerland; Bern, Switzerland. (sup c)DNA was isolated, digested with either EcoRI or PvuII, electrophoresed, blotted, and probed for DNA. Three distinct ribopatterns emerged, each with eight bands: patterns A, B, and C. Many (10 [59%] of 17) C. diphtheriae isolates were from skin or subcutaneous infections (wounds, ulcers) in Swiss patients and were found with Staphylococcus aureus, Streptococcus pyogenes, group C/G streptococci, or (sometimes) with gram-negative rods. Most patients in this subgroup were injecting drug users, homeless persons, prisoners, and (with one exception) men with a mean age of 40 (30 to 58) years. Another subgroup (from Germany) consisted of six patients with joint or bone infections (mentioned briefly in an earlier publication on coryneform bacteria from such infections; a seventh patient had C. diphtheriae biotype gravis [7]). The six had been hospitalized at the Endo Clinic in Hamburg, which specializes in orthopedic surgery. Four had had implantations of hip endoprostheses and had been admitted for prosthetic infections (with coagulase-negative staphylococci or S. aureus); one pure culture of C. diphtheriae biotype mitis was obtained from a patient with a knee prosthesis, and one mixed culture with Peptostreptococcus magnus and Peptostreptococcus prevotii was obtained from a patient who had a purulent knee infection after a fracture. Their average age was 60 (23 to 82) years and, to our knowledge, none was a drug user. While the Hamburg isolates were recovered only once from every patient, their isolation dates stretched from March 1, 1995, until July 26, 1996. Three were isolated on the day of admission, and all of them were isolated by different technologists. Finally, one child (from France) had endocarditis with C. diphtheriae biotype mitis. All isolates had identical antimicrobial susceptibility patterns. They were susceptible to amoxicillin (MIC, 0.25 µg/ml), amoxicillin-clavulanic acid (0.06 µg/ml), chloramphenicol (2 µg/ml), ciprofloxacin (0.25 µg/ml), clarithromycin (0.03 µg/ml), clindamycin (0.25 µg/ml), imipenem (0.03 µg/ml), penicillin (0.25 µg/ml), and vancomycin (1 µg/ml). In contrast, all strains were resistant to tetracycline in the disk diffusion test (inhibition zone diameter 10 mm to 11 mm); their MICs for tetracycline, doxycycline, and minocycline were 64 µg/ml, 16 µg/ml, and 16 µg/ml, respectively. While this type of isolated tetracycline resistance was typical for the isolates from Swiss injecting drug users (3), tetracycline resistance in nontoxigenic C. diphtheriae has otherwise very rarely been observed in Europe (8). It has, however, been reported in toxigenic C. diphtheriae isolates from Indonesia and, rarely, from Canada (9). The mechanism conferring this resistance in our strains has not been investigated. On analysis with restriction enzyme Pvu II, three different ribopatterns with eight bands each were found among the strains (Figure). These patterns, though distinct, shared five (A vs. C, B vs. C) and six (B vs. A) bands, respectively, and, therefore, may be considered related, as they would be if criteria for pulsed-field gel electrophoresis were applied (10). This view is supported by the fact that all strains had identical EcoRI patterns (not shown). All Swiss isolates were identical with the use of both enzymes, whereas both the Swiss pattern and a second pattern were found among the Hamburg isolates. The third pattern was found exclusively in the single French strain. The staphylococcal and streptococcal strains were not typed; they are no longer available to us. Our isolates thus resemble those Figure. Ribosomal RNA gene restriction patterns reported from obtained by using restriction enzyme Pvu II of Switzerland between Corynebacterium diphtheriae isolates belonging to 1990 and early 1996 ribotypes B (lanes 1 to 3), A (lanes 4 to 6 and 8), (2,3), which were and C (lane 7). also often accompanied by S. aureus or beta-hemolytic streptococci. Such nontoxigenic C. diphtheriae mitis may cause endocarditis, arthritis, and osteomyelitis (11,12). Most of the 52 isolates from France (11), examined with restriction enzymes different from ours and not available to us, also belonged to one ribotype. Their relatedness to our strains is unknown; however, they were largely tetracycline-susceptible. The two throat isolates from St. Petersburg, Russia, associated with a fatal diphtherialike disease (12) were not typed or tested for antibiotic susceptibility. The origin of the isolates we describe is unknown. They may have been present (but unrecognized) in the population for a long time. The mode of transmission is most likely common use of drug paraphernalia in the injecting drug use cases and in the endocarditis case; transmission is very difficult to explain in the orthopedic infection subgroup. Although the Swiss and the Danish borders are not close, migration and contacts are not uncommon among injecting drug users. The isolates may also have been distributed through the drugs themselves, as recently reported for S. pyogenes in Switzerland (13). Our study may be representative for Switzerland and Germany, but since submitting C. diphtheriae strains to a central reference laboratory in these countries is not mandatory, we cannot estimate the frequency of these strains. These strains may also have spread to other European countries; this hypothesis can only be tested in a large multicenter study of European diphtheria reference laboratories. --------------------------------------------------------------------------- Acknowledgments We thank J. Lüthy-Hottenstein and V. Pünter-Streit for excellent technical assistance and P. Das for her help in securing the strains from Hamburg. G. Funke is recipient of a European Society for Clinical Microbiology and Infectious Diseases research fellowship. Dr. Funke is director of clinical microbiology at Gärtner & Colleagues Laboratories in Weingarten, Germany. His areas of expertise are clinical and systematic bacteriology. Research interests include taxonomy and disease associations of coryneform bacteria and actinomycetes, as well as rapid methods for identification and susceptibility testing of medically relevant bacteria. Address for correspondence: Alexander von Graevenitz, Department of Medical Microbiology, University of Zurich, Gloriastra sse 32, CH-8028 Zurich, Switzerland; fax: 411-634-4906; e-mail: avg@immv.unizh.ch. References 1. Sloss JM, Hunjan RS. Incidence of non-toxigenic corynebacteria diphtheria in British military personnel in Germany. J Infect 1996;33:139. 2. Gubler J, Huber-Schneider C, Gruner E, Altwegg M. An outbreak of non-toxigenic Corynebacterium diphtheriae infection: single bacterial clone causing invasive infection among Swiss drug users. Clin Infect Dis 1998;27:1295-8. 3. Gruner E, Zuber PLF, Martinetti-Lucchini G, von Graevenitz A, Altwegg M. A cluster of non-toxigenic Corynebacterium diphtheriae infections among Swiss intravenous drug abusers. Medical Microbiology Letters 1992;1:160-7. 4. Von Graevenitz A, Funke G. An identification scheme for rapidly and aerobically growing Gram-positive rods. Zentralbl Bakteriol 1996;284:246-54. 5. Martinetti-Lucchini G, Gruner E, Altwegg M. Rapid detection of diphtheria toxin by the polymerase chain reaction. Medical Microbiology Letters 1992;1:276-83. 6. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing; eighth informational supplement [NCCLS document M 100-S8]. Wayne (PA): The Committee; 1998. 7. von Graevenitz A, Frommelt L, Pünter-Streit V, Funke G. Diversity of coryneforms found in infections following prosthetic joint insertion and open fractures. Infection 1998;26:36-8. 8. Patey O, Bimet F, Emond JP, Estrangin E, Riegel P, Halioua B, et al. Antibiotic susceptibilities of 38 non-toxigenic strains of Corynebacterium diphtheriae. J Antimicrob Chemother 1995;36:1108-10. 9. Rockhill RC, Sumarmo, Hadiputranto H, Siregar SP, Muslihun B. Tetracycline resistance of Corynebacterium diphtheriae isolated from diphtheria patients in Jakarta, Indonesia. Antimicrob Agents Chemother 1982;21:842-3. 10. Tenover FC, Arbeit RD, Goering RV, Mickelsen PA, Murray BE, Persing DH, et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995;33:2233-9. 11. Patey O, Bimet F, Riegel P, Halioua B, Emond JP, Estrangin E, et al. Clinical and molecular study of Corynebacterium diphtheriae systemic infections in France. J Clin Microbiol 1997;35:441-5. 12. Rakhmanova AG, Lumio J, Groundstroem KWE, Taits BM, Zinserling VA, Kadyrova SN, et al. Fatal respiratory tract diphtheria apparently caused by nontoxigenic strains of Corynebacterium diphtheriae. Eur J Clin Microbiol Infect Dis 1997;16:816-20. 13. Streptokokken-Infektionen bei Drogensüchtigen in der Region Bern. Bulletin des Bundesamtes für Gesundheit 1997;44:3. ______________________________________________________________________ Letters ______________________________________________________________________ Letters First Case of Human Ehrlichiosis in Mexico To the Editor: Ehrlichiosis is a zoonotic disease transmitted to humans through the bite of infected ticks (1). The first recognized human ehrlichial infection, Sennetsu fever, was described in Japan in 1954 (2). The first case of human ehrlichiosis in the United States was recognized in 1986 and was reported in 1987 (3). The disease is caused by intracellular gram-negative bacteria of the Ehrlichia genus. The bacteria can be found in the monocytes and granulocytes of peripheral blood. Human monocytic ehrlichiosis is caused by E. chafeensis, and human granulocytic ehrlichiosis is caused by E. equi or E. phagocytophilia, which was first recognized in 1994 (4). Most cases occur between April and September, and the reservoirs are field animals such as rodents, deer, and dogs. The clinical spectrum of the disease is similar to that of other febrile illnesses; without adequate and timely treatment, approximately 5% of the patients die (5). In the United States, more than 400 cases of serologically confirmed E. chaffensis infection have been documented since 1996 (6). No cases have been reported in Mexico. In February 1997, we evaluated a 41-year-old male patient from Merida. The patient had been exposed to ticks during activity in a rural area 1 week before the onset of illness. Clinical manifestations included frequent hyperthermia, rash, myalgia, headache, anorexia, fatigue, and cough. Physical examination showed bilateral cervical lymphadenopathy, and a chest radiograph showed an interstitial bilateral infiltrate. Hematic cytometry showed thrombocytopenia of 134 x 10(sup 3)/µL and 3200 leukocytes (1440 neutrophils/µL). Hepatic transaminases were elevated, with an aspartate aminotransferase: 92 U/L (normal: 22 U/L), alanine aminotransferase: 48 U/L (normal: 18 U/L), gamma-glutamyltranspeptidase: 278 U/L (normal: 28 U/L); and globulins: 4.8 g/dL with a polyclonal pattern. No antibodies against rickettsia, dengue virus, B-19 parvovirus, or HIV were detected. A serum sample gave a positive reaction by indirect immunofluorescence assay against E. chaffeensis at titers of 1:64 on week 2 and 1:128 on week 3. No infected monocytes or granulocytes were observed in peripheral blood. Remission of the clinical manifestations began on week 4 and was completed on week 6. This case indicates the existence of human ehrlichiosis in Yucatan, Mexico. Reactivity to E. chaffeensis suggests human monocytic ehrlichiosis; however, as antibody testing was not performed with E. phagocytophila or E. equi, the possibility of human granulocytic ehrlichiosis cannot be excluded. In any event, case reports indicate the need for deliberate search for cases. Dengue is endemic in this area of Mexico, and ehrlichiosis should be considered as a differential diagnosis. Renán A. Gongóra-Biachi,* Jorge Zavala-Velázquez, † Carlos José Castro-Sansores,* and Pedro González-Martínez* *Centro de Investigaciones Regionales "Dr. Hideyo Noguchi," Mérida, Yucatán, México; and †Facultad de Medicina, Universidad Autónoma de Yucatán, Mérida, Yucatán, México References 1. Dumler SJ, Bakken JS. Ehrlichial diseases of humans: emerging tick-borne infections. Clin Infect Dis 1995;20:1102-10. 2. Schaffner W, Standaert SM. Ehrlichiosis—in pursuit of an emerging infection. N Engl J Med 1996;334:262-3. 3. Maeda K, Markowitz N, Hawley RC, Ristic M, Cox D, McDade JE. Human infection with Ehrlichia canis, a leucocytic rickettsia. N Engl J Med 1987;316:853-6. 4. Bakken JS, Dumler JS, Chen SM, Eckman MR, Van Etta LL, Walker DH. Human granulocytic ehrlichiosis in the upper midwest United States: a new species emerging? JAMA 1994;272:212-8. 5. Walker D, Raoult D, Brouqui P, Marrie T. Rickettsial diseases. In: Fauci AS, Braunwald E, Isselbacher KJ, Wilson JD, Martin JB, Kasper DL, et al., editors. Harrison's principles of internal medicine. 14th ed. New York: The McGraw-Hill Companies; 1998. p. 1045-52. 6. Walker DH, Dumler JS. Emergence of the ehrlichioses as human health problems. Emerg Infect Dis 1996;2:18-29. HIV-1 Subtype F in Single and Dual Infections in Puerto Rico: A Potential Sentinel Site for Monitoring Novel Genetic HIV Variants in North America To the Editor: Although international efforts to systematically collect, characterize, and classify HIV isolates from around the world have increased considerably, data on HIV-1 genetic variations in Puerto Rico are limited. This island (population 3.7 million) has one of the highest AIDS incidence rates in the United States (53.3 cases per 100,000) (1). To evaluate the potential for a multiple subtype distribution pattern in Puerto Rico, we analyzed genetic variations between HIV-1 strains isolated from peripheral blood mononuclear cells of 63 asymptomatic HIV-infected female commercial sex workers from 12 communities. These participants were part of 290 female commercial sex workers followed in a larger cross-sectional study of risk behavior (2). HIV-1 subtypes F (n = 4) and B (n = 44) strains were identified in persons infected with a single viral subtype with a molecular screening assay based on restriction fragment length polymorphism (RFLP) analysis and with DNA sequencing of the viral protease gene-prot (3). The remaining 15 specimens were classified by RFLP as potential dual infections. Further cloning and sequencing of prot from three of these specimens confirmed one dual infection involving subtypes F and B viruses and identified two infections caused by genetically distinct quasispecies of subtype B variants. In further detailed pairwise analysis of HIV-1 prot genes, a small nucleotide divergence of 0.3% (0.0 to 1.1) within subtype F contrasted with a typical value of 6.3% (5.1 to 7.8) for the intrasubtype distance within subtype B prot sequences (4). The 99% similarity between prot subtype F Puerto Rican sequences suggested an epidemiologic link or a recent introduction of subtype F in Puerto Rico. Comparative sequence analysis of the C2-V3 env is useful in establishing the time that elapsed from infection on the basis of an annual nucleotide divergence of 0.5% to 1% in this region (5). Such analysis has been used to study the epidemiologic link between cases (4,6). Thus, we compared env sequences from two of five persons infected with prot subtype F strains. This analysis provided several observations. Env nucleotide divergence of 13.2% did not support a direct epidemiologic link between these strains. Furthermore, the relatively high intrasubtype diversity between env sequences suggested that evolution from a common progenitor would have taken a minimum of approximately 13 years. Phylogenetic analysis classified these two env sequences as subtype B, indicating that at least some of Puerto Rican prot subtype F viruses represent HIV-1 mosaics involving closely related prot F and significantly divergent env B sequences. Overall, discrepancy in both subtype assignment and nucleotide diversities within prot and env regions may indicate that distinct F/B mosaics circulating in Puerto Rico were likely the result of recombination between highly homogeneous subtype F of relatively recent arrival and divergent resident subtype B viruses. HIV-1 infections with subtype F strains including B/F mosaics have been reported in Brazil (3,7). To evaluate a potential HIV-1 linkage between Brazil and Puerto Rico, a comparative phylogenetic analysis was done on subtype F viral prot sequences from these countries. This analysis documented that HIV-1 subtype F strains in Puerto Rico are distinct from both Brazilian and Romanian viruses. Furthermore, our results show that genetic analysis of prot allows tracking of subtype F viruses of different origin. Recently, by this approach, HIV-1 prot subtype F of Puerto Rican origin and F prot/B env mosaic were identified in HIV-1-infected persons in New York city (8). Observation of HIV-1 subtype F strains in Puerto Rico together with the recent report describing the first cases of such infections in New York indicates the potential for further emergence of subtype F on the North American continent. The presence of a complex distribution pattern of subtype F infections in Puerto Rico has serious implications for the evaluation and development of HIV diagnostics and vaccines. Supported in part by grant G12RR-03050 (Y.Y.). The nucleotide HIV-1 sequences obtained in this study were submitted to GenBank; their accession numbers are AF096813-AF096833. Idhaliz Flores,* Danuta Pieniazek,† Nitza Morán,* Angel Soler,* Nayra Rodríguez,* Margarita Alegría,‡ Mildred Vera,‡ Luiz M. Janini,† Claudiu I. Bandea,† Artur Ramos,† Mark Rayfield,† and Yasuhiro Yamamura* *Ponce School of Medicine, Ponce, Puerto Rico; †Centers for Disease Control and Prevention, Atlanta, GA; ‡University of Puerto Rico, San Juan, Puerto Rico References 1. Centers for Disease Control and Prevention. HIV/AIDS surveillance report no. 2; 1997;9. 2. Drugs, HIV infection and risk behaviors among Puerto Rican sex workers, 1994-1996. Grant: NIAID/RCMI #G12RR03051 [Dr. Margarita Alegria]. Sociomedical Research Department, Graduate School of Public Health, University of Puerto Rico, Medical Sciences Campus, San Juan, Puerto Rico. 3. Ramos A, Tanuri A, Schechter M, Rayfield MA, Hu DJ, Cabral MC, et al. HIV-1 dual infections are an integral part of the HIV epidemic in Brazil. Emerg Infect Dis 1999;5:65-74. 4. Janini LM, Tanuri A, Schechter M, Peralta JM, Vicente AC, De la Torre N, et al. Horizontal and vertical transmission of human immunodeficiency virus type 1 dual infections caused by viruses of subtypes B and C. J Infect Dis 1998;177:227-31. 5. Myers G, Korber B, Berzofski JA, Smith RF, and Database and Analysis Staff, editors. Human retroviruses and AIDS 1991: a compilation and analysis of nucleic acid and amino acid sequences. Los Alamos (NM): Los Alamos National Laboratory; 1991. 6. Ou CY, Ciesielski C, Myers G, Bandea CI, Luo CC, Korber BT, et al. Molecular epidemiology of HIV transmission in a dental practice. Science 1992;256:1167-71. 7. Sabino EC, Shpaer EG, Morgado MG, Korber BT, Diaz RS, Bongertz V, et al. Identification of human immunodeficiency virus type 1 envelope genes recombinant between subtypes B and F in two epidemiologically linked individuals from Brazil. J Virol 1994;68:6340-6. 8. Weidle PJ, Ganea CE, Pieniazek D, Ramos CA, Ernst JA, McGowan JP, et al. Prevalence of HIV-1 group M, non-B-subtypes in Bronx, New York community: a sentinel site for monitoring of HIV genetic diversity in the United States. In: Proceedings of the 12th World AIDS Conference; 1998 Jun; Geneva, Switzerland [abstract no. 13225]. Paratyphoid Fever in India: An Emerging Problem To the Editor: Enteric fever is a major public health problem in India, accounting for more than 300,000 cases per year, Salmonella typhi is the most common etiologic agent (1), but Salmonella paratyphi A, the other causative agent, causes more asymptomatic infections than S. typhi. According to earlier reports from India, S. paratyphi A was implicated as a causative agent in 3%-17% of enteric fever cases (2). However, a large community-based study in an urban slum of Delhi during October 1995 to October 1996 found that S. paratyphi A caused approximately 20%-25% of the cases of enteric fever in this region (3). An outbreak of enteric fever due to a single S. paratyphi A strain in an urban residential area was reported in 1996 from New Delhi, where contaminated water was implicated as the probable source (4,5). This outbreak prompted a retrospective analysis of the laboratory records of the All India Institute of Medical Sciences, New Delhi, over a 5-year period (1994-1998) to study the change, if any, in the etiology of enteric fever in North India. We evaluated all blood culture records from the institute's clinical bacteriology laboratory for April to October (the months with the highest number of enteric fever cases) each year. Records were from patients residing in New Delhi and the surrounding areas of North India. The blood was collected by a phlebotomist in the outpatient department or by a resident doctor in hospital wards. Blood cultures were carried out by standard laboratory technique (6). Five ml of blood was added to 50 ml of brain heart infusion broth (Hi-Media Laboratory, India) under aseptic conditions. Bacterial identification was accomplished by standard microbiologic protocol (6). Susceptibility to antibiotics (amoxycillin, chloramphenicol, cotrimoxazole, gentamicin, ciprofloxacin, and ceftriaxone) was tested by the comparative disk diffusion method (Stokes method) (7). Chi-square for trend was calculated, and the p value was determined. The total number of blood cultures performed for enteric fever cases (10,109 in 1994, 12,092 in 1995, 17,652 in 1996, 15,997 in 1997, and 17,012 in 1998) did not change significantly over this period. The isolation of S. typhi changed little (Chi-square = 2.367; p = 0.123; statistically not significant). However, the proportion of S. paratyphi A isolates rose from 6.5% in 1994 to 44.9% in 1998 (Chi-square = 22.20; p <0.001; statistically significant). The proportion of S. paratyphi A isolations in enteric fever cases from 1994 to 1998 was 6.5%, 21.2%, 50.5%, 30.7%, and 44.9%, respectively. Even excluding the strains from the 1996 outbreak (4), we found that the proportion of S. paratyphi A in enteric fever cases increased compared with S. typhi (Chi-square = 30.528; p <0.001). With our catchment area, case definition of enteric fever, and laboratory methods remaining the same during this period, it appears that the etiology of enteric fever in North India is changing significantly. The age-wise distribution of S. typhi and S. paratyphi A showed that S. typhi was a significant isolate from children < 5 years of age, while this distribution was not observed for S. parathyphi A, which involved those > 5 years of age. Sex was not significantly associated (mean male to female sex ratio was 32.4:18 for S. typhi and 15.8:10.6 for S. paratyphi A). S. typhi has become increasingly sensitive to amoxycillin, chloramphenicol, and gentamicin, increasing from 75.1% in 1994 to 96.6% in 1998 for amoxycillin, from 71.9% in 1994 to 91.6% in 1998 for chloramphenicol, and from 96.4% to 100% for gentamicin. S. paratyphi A strains have remained uniformly sensitive (100%) to all antibiotics (amoxycillin, chloramphenicol, and gentamicin, as well as ciprofloxacin and ceftriaxone) used in the treatment of enteric fever. In light of reports of multidrug resistance in S. typhi, especially to quinolones, continued surveillance and monitoring of antimicrobial sensitivity of S. paratyphi A strains are needed. The increase in proportion of S. paratyphi A cases, which may be due to a high degree of clinical suspicion (with mild fever cases investigated for enteric fever), changing host susceptibility, or even change in the virulence of the organism, should be further investigated. Seema Sood, Arti Kapil, Nihar Dash, Bimal K. Das, Vikas Goel, and Pradeep Seth All India Institute of Medical Sciences, New Delhi, India References 1. Richens J. Typhoid and paratyphoid fevers. In: Oxford textbook of medicine. Weatherall DJ, Ledingham JGG, Warrell DA, editors. Vol 1. 3rd ed. London:: Oxford Medical Publication; 1996. p. 560-8. 2. Saxena SN, Sen R. Salmonella paratyphi A infection in India: incidence and phage types. Trans Royal Soc Trop Med Hyg 1966;603:409-11. 3. Kumar R, Sazawal S, Sinha A, Sood S, Bhan MK. Typhoid fever: contemporary issues as related to the disease in India. Round Table Conference Series on Water Borne Diseases. 12th ed. Ranbaxy Science Foundation, New Delhi, 1997;2:31-6. 4. Kapil A, Sood S, Reddaiah VP, Das BK, Seth P. Partyphoid fever due to Salmonella enterica serotype paratyphi A. Emerg Infect Dis 1997;3:407. 5. Thong K, Nair S, Chaudhry R, Seth P, Kapil A, Kumar D, et al. Molecular analysis of Salmonella paratyphi A from an outbreak in New Delhi, India. Emerg Infect Dis 1998;4:507-8. 6. Collee JG, Duguid JP, Fraser AG, Marmion BP. Mackie and Mc Cartney practical medical microbiology: laboratory strategy in the diagnosis of infective syndromes. 13th ed. London (UK): Churchill Livingstone; 1989. 601-7. 7. Stokes EJ, Ridgway GL. Clinical bacteriology: anti-bacterial drugs. 5th ed. London: Edward Arnold; 1980. p. 205-19. Hepatitis C Virus RNA Viremia in Central Africa To the Editor: Epidemiologic serosurveys have demonstrated high prevalence (6%-15%) of hepatitis C virus (HCV) infection in adults in sub-Saharan Africa (1-4). Although possible false-positive HCV serologic test results have been reported in Africa, HCV prevalence rates suggest a high rate of chronic infection among persons with anti-HCV antibodies (5,6). We have focused on HCV RNA infectivity of blood from donors attending the National Blood Center in Bangui, Central African Republic. We prospectively tested all blood donors between February and April 1998 for serum anti-HCV antibodies by both an HCV third-generation enzyme-linked immunosorbent assay (ELISA) (Abbott HCV EIA 3.0 test, Abbott, Chicago, IL, USA), which was chosen as a reference test for immunoglobulin (Ig) G antibodies to HCV, and by a simple membrane immunoassay system (Ortho HCV Ab Quik Pack, Ortho Diagnostic Systems Inc., Tokyo, Japan) (7). Anti-HCV-positive serum samples were further subjected to qualitative detection of HCV RNA by reverse transcription-polymerase chain reaction (AMPLICOR-HCV, Roche Diagnostic Systems, Inc., Branchburg, NJ, USA) (8). Of 163 serum samples (mean age ± standard deviation, 30±8 years), 155 were from male blood donors, 83 (51%) from first-time donors, and 125 (77%) from donors in the recipient's family. Fifteen (9.2%; 95% confidence interval [CI] 5%-15%) samples contained IgG to HCV by ELISA. Of the ELISA-positive samples, 14 were positive by the Quik Pack assay (sensitivity, 93.0%); of the 148 remaining ELISA-negative samples, 147 were negative by the Quik Pack assay (specificity, 99.3%). The agreement between the results of the two methods was 98.7%. Of the 163 samples, 10 (6.1%; CI 95%: 3%-11%) were positive for HCV antibodies (by ELISA and rapid test) and for HCV RNA. We confirmed a high prevalence of HCV-seropositivity among blood donors in Bangui and the subsequent high rate of HCV RNA viremic blood donations. To offset the major risk for transfusion-acquired HCV in Central Africa we recommend screening donated blood for anti-HCV. When laboratory facilities to perform ELISA are not available, the Quik Pack system, a simple reliable method for detecting anti-HCV antibodies in human serum that requires neither complex reagent preparation nor expensive instrumentation, could prove useful. Nicole Cancré,* Gérard Grésenguet,† François-Xavier Mbopi-Kéou, ‡Alain Kozemaka,† Ali Si Mohamed,* Mathieu Matta,* Jean-Jacques Fournel,§ and Laurent Bélec* *Université Pierre et Marie Curie, Hôpital Broussais, Paris, France; †Centre National de Transfusion Sanguine, Bangui, République Centrafricaine; ‡London School of Hygiene and Tropical Medicine, London, United-Kingdom; and §Hôpital de la Pitié-Salpêtrière, Paris, France References 1. Ndumbe PM, Skalsky J. Hepatitis C virus infection in different populations in Cameroon. Scand J Infect Dis 1993;25:689-92. 2. Xu LZ, Larzul D, Delaporte E, Bréchot C, Kremsdorf D. Hepatitis C virus genotype 4 highly prevalent in Central Africa (Gabon). J Gen Virol 1994;75:2393-8. 3. Fretz C, Jeannel D, Stuyver L, Herve V, Lunel F, Boudifa A, et al. HCV Infection in a rural population of Central African Republic (CAR): evidence for three additional subtypes of genotype 4. J Med Virol 1995;47:435-7. 4. Pawlotsky JM, Bélec L, Grésenguet G, Desforges L, Bouvier M, Duval J, et al. High prevalence of hepatitis B, C and E markers in young sexually active adults from the Central African Republic. J Med Virol 1995;46:269-73. 5. Aceti A, Taliani D. Hepatitis C virus testing in African sera. Ann Intern Med 1992;116:427. 6. Callahan JD, Constantine NT, Kataaha P, Zhang X, Hyams KC, Bansal J. Second generation hepatitis C virus assays: performance when testing African sera. J Med Virol 1993;41:35-8. 7. Kodama T, Ichiyama S, Sato K, Nada T, Nakashima N. Evaluation of a membrane filter assay system, Ortho HCV Ab Quik Pack, for detection of anti-hepatitis C virus antibody. J Clin Microbiol 1998;36:1439-40. 8. Young KKY, R. Resnick RM, Myers TW. Detection of hepatitis C virus RNA by a combined reverse transcription-polymerase chain reaction assay. J Clin Microbiol 1993;31:882-6. Immunization of Peacekeeping Forces(sup 1) To the Editor: The immunization status of military contingents arriving from different nations for peacekeeping missions may vary widely. This variation results from lack of information, coordination, and financial support. For larger missions, the United Nations (UN) Headquarters issues recommendations about needed vaccines; recently, operations officers have consulted World Health Organization experts before issuing recommendations, and their advice, which takes into account epidemiologic data in the host country, has improved. Medical officers who develop recommendations for smaller missions must consider the pathogenic agent; environment; host efficacy, safety, and price of preventive measures; and legal and ethical aspects. Data on the incidence of vaccine-preventable diseases within a military population that had similar duties in the same location are rarely available. When data from the respective region are not available, disease incidence or prevalence in the host country may be substituted. These data, however, may be misleading since the military often does not have the same lifestyle as the native population. Plague, for instance, had an incidence rate of 8 per 100,000 in Namibia, but not a single case was reported in the South African Armed Forces (unpub. SAMS report: Disease Profile of South West Africa, 1989). If epidemiologic documentation for a host country is not available, data from neighboring countries may be useful. Traveler's diarrhea is the most frequent health problem abroad (1,2). Although the diarrhea is self-limited and lasts an average of 1 day with appropriate treatment (4 days without), the unproductive time may be detrimental to a military mission. Oral vaccines against the three most frequent causes of traveler's diarrhea (enterotoxigenic Escherichia coli, Campylobacter spp., and rotavirus [1,2]) are being developed; the latter will be available soon (3). Hepatitis A, most frequent among the vaccine-preventable diseases (4), is 10 to 100 times more frequent than typhoid fever (4,5). Hepatitis B occurs mainly in expatriates, but infections have also been observed in tourists who have had unprotected casual sex (6). The incidence rate of rabies is unknown, but animal bites that may result in rabies virus transmission and thus necessitate postexposure prophylaxis are frequent (7). Only anecdotal cases of diphtheria, tetanus, and tuberculosis have been reported (8). Poliomyelitis, yellow fever, Japanese encephalitis, and plague occur only in limited parts of the world (5). The situation may rapidly change as epidemics occur (e.g., diphtheria in eastern Europe in the early and mid-1990s) (9). If needed, the World Health Organization can provide information on confirmed and unconfirmed epidemics on a weekly basis. Travel and peacekeeping mission statistics share similarities. In Namibia, the South African Armed Forces had most often observed hepatitis (unspecified), with rare cases of tuberculosis, typhoid, and meningitis (unpub. SAMS report: Disease Profile of South West Africa, 1989), as did the UN mission to Namibia, where within 12 months and with 7,114 employees, seven cases of hepatitis (mostly hepatitis A, some unspecified) occurred (10). No other vaccine-preventable infections were diagnosed in this UN mission. Considering both risk (on the basis of incidence rates) and impact of infection, the priority for immunization (from highest to lowest) is as follows: hepatitis A, hepatitis B, rabies, poliomyelitis, yellow fever, typhoid fever, influenza, diphtheria, tetanus, meningococcal disease, Japanese encephalitis, cholera, and measles. To administer all vaccines would be extremely costly and may also result in an increased rate of adverse side-effects. Immunizations against the more frequent, more severe infections should be given priority. If a mission is limited to one season, environmental factors of that respective season should be considered. This general rule is more important for vector-borne than for vaccine-preventable infections, except for influenza and meningococcal disease. Persons who are already immune (because of previous immunization or immunity after infection) need not be vaccinated. The latter cause is particularly often true of hepatitis A; troops recruited in developing countries have an anti-hepatitis A virus seroprevalence rate close to 100% (11). Hepatitis B immunization, except for non- and low-responders, probably grants lifelong protection (12); the same is likely for measles vaccine. Sometimes the host country may require proof of some specific vaccination based on the International Health Regulations (13), currently under fundamental revision to become a more effective tool in preventing the spread of infections that may be a global hazard (14). In addition to adequate epidemiologic information and coordination between the military, international health organizations, and the host country, successful intervention efforts require thorough knowledge of vaccine characteristics with varying rates of efficacy and duration of protection. Cost-benefit evaluations, which would be very desirable, are unlikely in areas of political instability. Robert Steffen Institute for Social and Preventive Medicine of the University, Zurich, Switzerland (sup 1)Presented in part at the NATO Research & Technology Organization, Aerospace Medical Panel Symposium on Aeromedical Support Issues in Contingency Operations, Rotterdam, The Netherlands, 1 October 1997. References 1. DuPont HL, Ericsson C. Prevention and treatment of travelers' diarrhea. Drug Therapy 1993;328:1821-7. 2. Farthing MJG, DuPont HL, Guandalini S, Keusch GT, Steffen R. Treatment and prevention of travellers' diarrhoea. Gastroenterology International 1992;5:162-75. 3. Levine MM, Svennerholm A-M. Prioritization of vaccines to prevent enteric infections. In: DuPont HL, Steffen R, editors. Textbook of travel medicine. 1st ed. Hamilton: B.C. Becker Inc.; 1997. p. 370. 4. Steffen R, Kane MA, Shapiro CN, Schoellhorn JK, Van Damme P. Epidemiology and prevention of hepatitis A in travelers. JAMA 1994;272:885-9. 5. World Health Organization. International travel and health. Geneva: The Organization; 1999. 6. Steffen R. Risk of hepatitis B for travellers. Vaccine 1990;8:31-2. 7. Hatz CF, Bidaux JM, Eichenberger K, Mikulics U, Junghanss T. Circumstances and management of 72 animal bites among long-term residents in the tropics. Vaccine 1994;13:811-5. 8. Steffen R. Travel medicine prevention based on epidemiological data. Trans R Soc Trop Med Hyg 1991;85:156-62. 9. Hardy IRB, Dittmann S, Sutter RW. Current situation and control strategies for resurgence of diphtheria in newly independent states of the former Soviet Union. Lancet 1996;347:1739-44. 10. Steffen R, Desaules M, Nagel J, Vuillet F, Schubarth P, Jeanmaire C-H, et al. Epidemiological experience in the mission of the United Nations Transition Assistance Group (UNTAG) in Namibia. Bull World Health Organ 1992;70:129-33. 11. Centers for Disease Control and Prevention. Hepatitis A immunization. MMWR Morb Mortal Wkly Rep 1996;45(RR-15):7. 12. Hall AJ. Hepatitis B vaccination: protection for how long and against what. BMJ 1993;307:276-7. 13. World Health Organization. International health regulations. 3rd annotated ed. Geneva: The Organization; 1983. 14. World Health Organization. Revision of the international health regulations. Wkly Epidemiol Rec 1997;72:213-5. Sexually Transmitted Diseases in Ukraine To the Editor: With the political changes in eastern Europe in the last 10 years have come social and economic changes (1). Ukraine not only faces almost insurmountable problems as it tries to form a new government, it also faces many serious health issues including sexually transmitted diseases (STDs). Surveillance data from the Ukrainian STD Center from January 1, 1989, through December 31, 1995, were analyzed on the basis of reports received through 1997. In western Europe, the incidence of syphilis and gonorrhea declined from 1980 to 1991 to less than 2% per 100,000 persons for syphilis and less than 20% per 100,000 persons for gonorrhea. However, in Ukraine, since 1989, the notification rate of syphilis has skyrocketed—from 5 per 100,000 persons in 1990 to 170 in 1995. In some regions, this rate exceeds 220 cases per 100,000 persons. Moreover, cases among children younger than 14 years of age are also increasing. In 1995, the syphilis rate for persons older than 30 years of age was 170 per 100,000; 600 per 100,000 girls younger than 15 years of age; and 1,550 to 2,000 per 100,000 girls 15 to 16 years of age. The large number of girls with the disease is in part due to teenage prostitution (1). Most syphilis and gonorrhea cases are attributed to sexual transmission. Explanations of this phenomenon include the rapid growth of the sex industry, increasing numbers of homeless persons and refugees in Ukrainian cities, poor diagnostic facilities, punitive legislation that reduces the likelihood of going to treatment services, and limited or inadequate treatment (2). The Ukrainian government is reviewing its arrangements for the control of STDs, including HIV/AIDS, to identify clear objectives and priorities. Education and treatment would be effective in preventing the spread of STDs in Ukraine, but these measures are inadequately funded (3). Evaluation and risk reduction are also great weapons in preventing the spread of STDs (4). However, the response of the local and world communities has been inadequate in stemming a major STD epidemic in Ukraine. United Nation's Children's Fund (UNICEF) is developing a long-term program in Ukraine with a focus on STDs in adolescents and youth. This comprehensive program will tackle not only STDs but other related issues, such as HIV and teenagers' reproductive health (5). Greater coordination of the agencies responsible for STD control in Ukraine will be sought, together with an expansion of health promotion and prevention projects for young persons and groups at high risk (6). An effective strategy for the control of STDs in Ukraine will, therefore, need to find ways to modify current programs and the way they interact to create effective control interventions. Dmitry I. Ivanov University of Alabama at Birmingham, Birmingham, Alabama, USA References 1. Dittmann S, Gromyko A, Mikkelsen H, Schaumburg A, Adamian R, Khodakevich L, et al. Epidemic of sexually transmitted diseases in eastern Europe. Geneva: World Health Organization; 1996. 2. Kobyshcha Y. HIV risk-related behavior of homo and bisexual men and STD patients in Ukraine. National AIDS Committee and Center 1994;7:290-3. 3. Normand J, Vlahov D, Moses LE. Preventing HIV transmission: the role of sterile needles and bleach. The effects of needle exchange programs. Washington: National Academy Press; 1995. p. 208-55. 4. Spinhenko Y. Prevention of the spread of AIDS in the Ukrainian SSR. Lik Sprava 1988;9:1-3. 5. Usenko A, Grazhdanov N, Stepanets V, Neshcheret E, Maksiutenko E. Effective knowledge propaganda in the chief strategy for preventing HIV infection among adolescents. Lik Sprava 1994;9:192-6. 6. Tichonova L, Borisenko K, Ward H, Meheus A, Gromyko A, Renton A, et al. Epidemics of syphilis in the Russian Federation: trends, origins, and priorities for control. Lancet 1997;350:210-3. Yellow Fever Vaccine To the Editor: Monath et al. (1) outlined existing facilities for distribution of yellow fever vaccines in the United States and pointed to difficulties for prospective vaccinees in remote locations. Their recommendation that primary health-care providers be allowed to dispense yellow fever vaccination merits serious consideration. Acceptance of such a strategy in the United States would inevitably be emulated elsewhere. Nevertheless, before such a strategy is approved, vaccine potency should be monitored at distribution points, and a sample of vaccine recipients should be examined for vaccine-induced immune response. In Nigeria, systematic investigation of yellow fever vaccine distribution and transportation to remote locations has found loss in vaccine potency. Vaccine in storage sites and immunization centers in Lagos was fully potent, but potency in Osun and Oyo was 016 log(sub 10) to 0.22 log(sub 10) lower than the stipulated level (2). Furthermore, the titer of two vaccine lots that had been frozen after reconstitution from their lyophilized state dropped from the initial 3.15 log(sub 10) to 3.53 log(sub 10) to zero. If the United States were to implement an extended strategy, similar studies of vaccine lots should be conducted to determine whether every vaccinee has received a full dose of yellow fever vaccine. In Illinois during the early 1970s, weak links in maintenance of refrigeration facilities and use of outdated vaccines in vials exposed to the sun for long hours were reported for live poliovirus vaccines (3). In the Northern Territory of Australia, examination of 144 vials of hepatitis B vaccine formulations during transport to immunization centers showed that 47.5% had been exposed to temperatures of -3°C or lower (4). Assays of the potency of yellow fever vaccine, as well as quantification of vaccine-induced neutralizing antibody, is a multistep procedure that relies on inoculation of mice or Vero or polysaccharide cells (5). The successful "take" of yellow fever vaccine can be determined starting the second postvaccination day by demonstrable viremia detected by reverse-transcriptase polymerase chain reaction and by marked increases in neopterin, beta2-microglobulin, and circulating CD8+ cells (6). Alternatively, elevated levels of tumor necrosis factor and interleukin-1 receptor antagonists on day two after vaccination (7) could be used to monitor the success of vaccinations by primary-care providers in remote areas in the United States (1) and elsewhere. During the 1990s, isolation of yellow fever virus was reported in persons with a nonspecific febrile illness that did not meet the case definition of yellow fever (8). Air travel by such persons to the United States, which has areas infested by Aedes aegypti, could initiate yellow fever epidemics; because these travelers would have a nonspecific febrile illness, they would escape the existing surveillance network. In conclusion, introducing yellow fever immunizations by primary health-care providers would be ideal, only with a concurrent plan to monitor vaccine potency at immunization centers and obtain in vitro evidence of a successful vaccine take. Such a strategy would blunt yellow fever–associated deaths, illnesses, and symptomless viral carriage in the community. Subhash C. Arya Centre for Logistical Research and Innovation New Delhi, India References 1. Monath TP, Giesberg JA, Fierros EG. Does restricted distribution limit access and coverage of yellow fever vaccine in the United States? Emerg Infect Dis 1998:4:698-702. 2. Adu FD, Adedeji AA, Esan JS, Odusanya OG. Live viral vaccine potency: an index for assessing the cold chain system. Public Health 1996;110:325-30. 3. Rasmussen CM, Thomas CW, Mulrooney RJ, Morrissey RA. Inadequate poliovirus immunity levels in immunised Illinois children. Am J Dis Child 1973;126:465-9. 4. Miller NC, Harris MF. Are childhood immunization programmes in Australia at risk? Investigations of the cold chain in the Northern Territory. Bull WHO 1994;72:401-8. 5. World Health Organization. Techniques for potency evaluation of yellow fever vaccine. Technical Report Series 1998;872:67-8. 6. Reinhardt B, Jaspert R, Niedrig M, Kostner C, L'age-Stehr J. Development of viremia and humoral and cellular parameters of immune activation after vaccination with yellow fever virus strain 17D: a model of human flavivirus infection. J Med Virol 1998;56:159-67. 7. Hacker UT, Jelinek T, Erhardt S, Eigier A, Hartmann G, Nothdurft HD, et al. In vivo syntheses of tumor necrosis factor-alpha in healthy humans after live yellow fever vaccination. J Infect Dis 1998;177:774-8. 8. Sanders EJ, Maffin AA, Tukei PM, Kuria G, Ademba G, Agata NN, et al. First recorded outbreak of yellow fever in Kenya, 1992-1993. I. Epidemiologic investigations. Am J Trop Med Hyg 1998;59:644-9. Yellow Fever Vaccine—Reply to S. Arya To the Editor: Dr. Arya correctly points out that there have been problems with degradation of live viral vaccines, including yellow fever vaccines, that have not been properly handled and stored at the point of use. However, in the United States and western Europe, yellow fever vaccines are stabilized and require the same storage facilities at the point of use as other vaccines routinely distributed by family physicians and pediatricians. Varicella vaccine (and even measles vaccine) is less stable than yellow fever vaccine but is distributed to all registered physicians in the United States. Since vaccines and other perishable medicines are typically shipped by overnight courier services using qualified methods that ensure maintenance of low temperature, there is no barrier to use of a similar system for yellow fever vaccine. Empirical testing for antibody, viremia, or even surrogate markers of T-cell activation may be useful; however, it is difficult and expensive, involves unvalidated tests with unknown sensitivity and specificity, and is unnecessary, except under very special circumstances. A more direct measure of vaccine stability is direct potency measurement of samples stored at the point of use, as was done in the cited study in Nigeria by Adu et al. However, given the current controls on vaccine distribution in the United States, we do not believe that there would be a need to validate vaccine effectiveness at point of use in the event of a change of policy with respect to vaccinating centers. The cold-chain infrastructure and the training of medical personnel in vaccine storage and administration may not provide the same assurances in other countries. While our suggested changes to the system of yellow fever distribution may improve vaccine coverage and have other desirable benefits in the United States, they would not be appropriate for less stable systems for vaccine supply and use. T.P. Monath, J.A. Giesberg, and E.G. Fierros OraVax, Cambridge, Massachusetts, USA ________________________________________________________________________ News and Notes ________________________________________________________________________ News and Notes 54th International Northwestern Conference on Diseases in Nature Communicable to Man August 15-18, 1999 The 54th International Northwestern Conference on Diseases in Nature Communicable to Man at Utah State University, Logan, UT, is scheduled for August 15-18, 1999. Call for abstracts in the following general categories are proposed, but may be modified: viral zoonoses, bacterial zoonoses, water and foodborne infections, vector-borne diseases, and curiosa et exotica. The deadline for abstracts is April 15, 1999. The R.R. Parker Memorial Address will be presented by Jesse Goodman, Senior Advisor for Medical Programs, U.S. Food and Drug Administration. To obtain program announcements and registration forms, make hotel reservations, or submit abstracts, contact one of the following persons: Dhitinut Ratnapradipa, scientific program coordinator, e-mail: dratnapr@health.utah.edu; or Robert Elbel, meeting coordinator, e-mail: elbel@bioscience.utah.edu, tel:(801) 581-4816, fax: (801) 581-4668. Program information is also available at the following web site: http://www.mosquito.org./upcoming.html. ----------------------------------------------------------------------- Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: ftp://ftp.cdc.gov/pub/EID/vol5no3/ascii/vol5no3.txt Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; >/= for greater than or equal to.